Book - Chemical embryology 1 (1900) 3

From Embryology
Embryology - 9 Aug 2020    Facebook link Pinterest link Twitter link  Expand to Translate  
Google Translate - select your language from the list shown below (this will open a new external page)

العربية | català | 中文 | 中國傳統的 | français | Deutsche | עִברִית | हिंदी | bahasa Indonesia | italiano | 日本語 | 한국어 | မြန်မာ | Pilipino | Polskie | português | ਪੰਜਾਬੀ ਦੇ | Română | русский | Español | Swahili | Svensk | ไทย | Türkçe | اردو | ייִדיש | Tiếng Việt    These external translations are automated and may not be accurate. (More? About Translations)

A personal message from Dr Mark Hill (May 2020)  
Mark Hill.jpg
I have decided to take early retirement in September 2020. During the many years online I have received wonderful feedback from many readers, researchers and students interested in human embryology. I especially thank my research collaborators and contributors to the site. The good news is Embryology will remain online and I will continue my association with UNSW Australia. I look forward to updating and including the many exciting new discoveries in Embryology!

Needham J. Chemical Embryology Vol. 1. (1900)

Online Editor  
Mark Hill.jpg
This historic 1900 volume 1 of a textbook by Needham describes chemical embryology.

Internet Archive

Also by this author: Needham J. Chemical Embryology Vol. 2. (1900)

Modern Notes:
Historic Textbooks

Historic Disclaimer - information about historic embryology pages 
Mark Hill.jpg
Pages where the terms "Historic" (textbooks, papers, people, recommendations) appear on this site, and sections within pages where this disclaimer appears, indicate that the content and scientific understanding are specific to the time of publication. This means that while some scientific descriptions are still accurate, the terminology and interpretation of the developmental mechanisms reflect the understanding at the time of original publication and those of the preceding periods, these terms, interpretations and recommendations may not reflect our current scientific understanding.     (More? Embryology History | Historic Embryology Papers)

Part III General Chemical Embryology

All things began in order, so shall they end, and

so shall they begin again, according to the ordainer

of order and the mystical mathematicks of the city

of heaven.

Sir Thomas Browne.

Preliminary Note

There have already been certain reviews of work in chemical embryology as a whole, among which those of Paechtner and Schulz are the most valuable. The former dealt almost exclusively with the chemistry of the egg from a static viewpoint, and only devoted a short section to the metabolism of the embryo during its development, while the latter, though dealing specifically with embryonic metabolism, gave hardly more space to it than Paechtner. In both cases the discussion was little more than a catalogue of references, and in neither case was the literature anything like complete, including, indeed, less than a tenth of the relevant citations.

The first review of chemical embryology was written by Grafe in 19 10, but, though he outlined several valuable ideas, it is now of small importance. Good information may, however, be found in Aron's monograph on the chemistry of growth and on the mammalian side there are Harding and Murlin. Other, less satisfactory, reviews are by Cazzaniga and Steudel. Finally there is, of course, an immense amount of work which can be found in no review, for investigators have followed the counsel of Godlevski (1910): "Unsere Kenntnisse hinsichtlich der chemischen Zusammensetztung der Eier noch lange nicht ausreichend sind, so waren weitere Forschungen auf diesem Gebiete auch aus dem Grunde sehr erwunscht weil sie den Ausgangspunkt fiir die Physiologic des embryonalen Stoffwechsels welcher bisher gleichfalls nur sehr wenig untersucht wurde, bilden mussen".

Every effort has been made to give an accurate and complete presentation of the data in the Tables of this book and of the experimental conditions under which they were obtained, but investigators should always consult in addition, whenever possible, the relevant original memoirs referred to in the Bibliography.

Section 1 The Unfertilised Egg as a Physico-Chemical System

1-1. Introduction

In giving an account of the present state of our knowledge about the chemical constitution of the egg-cell and the food-material which is accumulated around it or inside it, I shall not follow a strictly logical order of exposition, according to the phyla of systematic biology. I have judged it best to begin with the egg of the hen, for not only is it the most familiar and the best known of all eggs, but it is also the one which has been most thoroughly investigated biochemically.

It should be remembered that the two main morphological divisions of the egg, (a) the egg-cell itself and {b) its coverings, appear in protean modifications throughout the animal kingdom. The former may be a simple cell with its ooplasm, nucleus, nucleolus, etc., as in the echinoderms, and no covering at all save its cell-membrane, or at the other extreme it may be swollen up with food-material or yolk to the prodigious proportions of the avian egg-cell. The membrane again may be a thin coat of investing cells such as the tunicate egg possesses, or it may be the jelly of the amphibian egg, or, again, it may be the complex arrangement of egg-white, chalazae, shellmembranes, and shell, which is present in the bird's egg. All imaginable degrees of richness in yolk are present in the egg-cells of animals, and upon this fact depend the various kinds of cleavage which they show: alecithic eggs, on the one hand, such as those of most invertebrates, having a holoblastic form of development in which the whole egg participates in cleavage; and yolk-rich eggs, on the other hand, such as those of most vertebrates, having a meroblastic development, only a localised part of the egg undergoing cleavage, the rest remaining as a sac full of yolk until it is finally absorbed.

1-2. General Characteristics of the Avian Egg

After the historical introduction which has been given, it should be unnecessary to remark on the general arrangement of the bird's egg. We have with Harvey referred to it as an exposed, and, as it were, detached uterus, and with Fabricius ab Aquapendente we have enumerated the parts of the typical avian ovum. Fig. 13, however, shows the general disposition of parts diagrammatically.

Fig. 13. Diagrammatic representation of the hen's egg. The chalazae were called by Tredern Ligamenta albuminis. Bartelmez gives a discussion of the factors governing the angle which the embryonic axis makes with the axis of the egg as a whole. The yolk is not a perfect sphere but lengthened along the main axis. The egg-white is divisible into three layers which increase in density from without inwards. The chalazae, as Berthold was the first to find, are not present in reptilian eggs.

First, as to size and shape. The size and shape of the egg were shown by Curtis in 191 1 and by Surface in 191 2 to be due partly to the structure of the oviduct, which very probably may be considered an inherited character, as was claimed by Newton. D'Arcy Thompson's discussion of the mechanics of egg-formation in birds, in his Growth and Form, will be famiHar, but some biologists, such as Horwood, have taken exception to his conclusions about the physical influences which shape the egg. Ernst's well-known experiment was the startingpoint of these discussions ; she caused hens to lay on a surface of wet sand and charcoal, and so, observing the process, found the blunt end to be blackened. This was in agreement with many other observers, such as V. Nathusius; Landois; Jasse; Konig-Warthausen and Erdmann; and d'Arcy Thompson accordingly described the hen's egg as moving down the oviduct blunt end forwards, the pointed end owing its form to the peristaltic compression of the oviduct. Unfortunately all observers agree (Purkinje; von Baer; Coste; Kiitter; Taschenberg; Wickmann and Patterson for the hen, Blount and Patterson for the pigeon, Kiitter for the hawk, and Wickmann for the canary) that the pointed end passes first down the oviduct. It appears that the egg must turn right round in the act of being laid, and Bartelmez, indeed, has seen this occur. Curtis has shown that the shape of the egg depends to some extent upon its size and this biometric observation was afterwards confirmed by Pearl & Curtis. Many abnormalities have been reported in eggs. They need merely be mentioned here with their authorities, thus :





Eggs containing masses of tissue^ more or less organised.

von Nathusius.



Dwarf eggs.

Pearl & Curtis.



Ovum in ovo.






Pearl & Curtis.



Roberts & Card.


Double and triple-yolked eggs.


Parker. '




Inadequate shell.

Riddle & King.

Dwarf or absent yolk (ovum centennium^).

Mercier. Szuman. Bugnion. Gelabert.

^ See Sir Thos. Browne, Pseudodoxia Epidemica, Bk iii, ch. 7, "Of the basilisk". The eggs of Chelonia also, according to Deraniyagala, are sometimes laid without yolks.

It is interesting in this connection that Riddle has traced the occasional production of eggs with deficiency of white and shell but not of yolk, to a lack of the thymus hormone which he has called "Thymovidine". Feeding with desiccated thymus removed completely these effects. "The whole of the data", he said, "seem to demonstrate the presence in the thymus of a substance having a highly specific action on the oviduct of birds — and presumably on that of all those vertebrate animals which secrete egg-envelopes." The syndrome involved eggs with normal yolks but hardly any shell or albumen, frequent reduction of normally paired ovulations to single ovulations, diminished fertility, and restricted hatchability of the eggs. "Though not necessary to the life of the individual", said Riddle, "thymovidine would seem to be essential to the perpetuation of those vertebrate species whose eggs are protected by egg-envelopes. Such animals were the ancestors of mammals and thus mammals could hardly have come into existence without the thymus." These considerations are of much interest in view of other speculations on the evolutionary aspect of chemical embryology, e.g. Section 6-6. They also suggest that the mammalian thymus is now a vestigial organ.

The air-space, the shell and the white of the normal egg need no special remark at present, but the yolk is a more complicated structure. Around a central core of "white" or "milky" yolk the yellow yolk is secreted in the ovary of the hen in concentric layers, which form the appearance of "haloes" in the finished egg, and which show up especially clearly when the hen is fed on Sudan III or some other nontoxic dye which has a selective staining action on fat. The white yolk in the centre is continued in a flask-like shape (the latebra) up to the surface of the yolk underneath the germinal disc, and is then continued in a very thin layer all round the exterior of the yolk underneath the vitelline membrane. The white yolk is thus the first nourishment of the embryo. It is not certain whether there are also layers of white yolk between the concentric layers of yellow yolk, for they have never been analysed chemically, and Balbiani maintains that they only differ from the yellow layers by having less yellow pigment. The differences between the true white yolk and the yellow yolk are, as will be seen later, far more profound. Balfour & Foster, in their Elements of Embryology of 1877, described the yellow yolk as consisting histologically of spheres of from 25 to loo/x in diameter, filled with numerous minute highly refractive granules and very susceptible to crushing and rough treatment. After boihng, the spheres assume a polyhedral form. The granules seen within them must consist of protein, for they are not soluble in ether or alcohol. On the other handjthe white yolk elements are vesicles smaller than the globules of the yellow yolk, being about 4 to 75 /n across, with a highly refractive body, often as small as i [x, in the interior of each. These vesicles are sometimes collected together into much larger vesicles. They observed also underneath the blastoderm or the germinal disc a number of large vacuoles filled with fluid — large enough, in fact, to be seen with the naked eye. The histology of yolk has been reviewed by Dubuisson, and at one time many papers were published on it, e.g. those of Virchow. They cannot be considered in detail here.

1-3. The Proportion of Parts in the Avian Egg

Of the weight of the whole egg, the shell takes up about 10 per cent., the albuminous white 50 per cent, and the yolk 30 per cent, in round numbers. These relationships have been determined by a multitude of investigators, whose results are drawn up in Table i .

Table i . Distribution of the parts in the egg,

Italic figures represent dry weight only.


egg weights










and date

Hen, Polish i



van Hamel-Roos ('.890)

,, Polish ii




,, Holland (Zwol.)




„ Holland (Tiel)









Miinster Ag. Sta. (1900)





Drechsler (1896)










Plimmer (1921)


I i-i








Langworthy (190 1-2)





















Lebbin (1900)





Welmanns (1903)






Segin (1906)






Liihrig (1904)







Hen (various breeds)





von Czadek (191 7)





Rose (1850)





Hen (various breeds)




Carpiaux (1903)




Lehmann (1850)




Prout (1859)




Poleck (1850)



Stained by Kossa's method for the detection of calcium phosphate. The considerable variations in the vitelline globules may be noted. Magnification, 6xD: prepared and microphotographed by Dr V. Marza.



SECT. l]



Table i {cont.)


egg weights










and date

Hen, Leghorn



Murray (1925)

Hen (various breeds)





Iljin (1917)









van Meurs (1923)












Voit (1877)

Nidicolous birds




Tarchanovf (1884)

Starling ...












Canary ...




Thrush ...















Nidifugous birds

















Turkey ...













17-78 (grains)




Glikin (1908)






Davy (1863)





















3 J

Golden-crested wren





) )








































(gm-) 57-57




Hartung (1902)





Voit (1881)



Fere (1896)









Pott & Preyer (1882)






Rozanov (1926)






Hepburn & Katz (1927)











Baudrimont & de St Ange (1846)

Dwarf hen ...




Sacc (1847)



Pott (1879)





Atwater & Bryant ( 1 906)


f- All Tarchanov's figures exclude the shell weight.

Weight of








tents (gm.)





and date


• 57-12





Friese (1923)


• 137-38






781 1






• 92-93


















• 25-40












. 27-03





Blackbird ...
















Weight of








ents (gm.)




and date

Plover ( Vanellus crist






Bauer (1893-5)

Hen {Gallus domestict







Guinea-fowl {Meleag

ris gallopavo)





Swallow (Hirundo ru.






Partridge [Perdrix ci






Sparrow {Passer dom



Thrush (Turdus

? ) ...





Duck (Anas) [doubl






The above data were all obtained without any ad hoc investigation of the probable errors involved in weighing eggs and parts of eggs. An elaborate study by M. R. Curtis in 1 9 1 1 gave the following results on Gallus domesticus :

Actual weight in gm. %

56-04 100

33-22 59-26

16-31 29-14

6-28 ii-i8

0-23 0-42

Whole egg Albumen


Shell and membranes Error

But though this is the case with the egg in its natural state, the solid matter is concentrated much more in the -yolk than in the white, so that, as the analyses of Poleck and Iljin, for instance, show, for dry weight the conditions are exactly reversed. The egg-white may, indeed, be regarded as the principal reservoir of water for the embryo which develops on dry land, and this is a point which will be discussed later (see Section 6-6). The eggs of different breeds of hen vary to some extent in the relative weights of shell, white and yolk; but, although it is difficult to lay down any general rule, these variations do not greatly exceed the variations due to factors connected with the individual hen. Iljin's lightest shells make up about 7 per cent, of the G,gg weight and the heaviest not more than 11-5 per cent.

It is certain that there are constant differences between the eggs of different breeds, but as a whole these are quite outweighed by individual differences, and only appear when extended statistical studies are undertaken. The eggs of other birds, however, do not fall within these limits. Langworthy, for example, has shown that, in the duck's egg, the shell may account for as much as 14 per cent, of the whole weight. A similar result was found for the turkey and the goose, while the guinea-fowl's egg has a shell of nearly 1 7 per cent, of the whole weight. The wide series of Friese, shown in Table i, seems to indicate that the larger the egg the more shell it has to have : thus the canary's egg weighing just under 2 gm. has 4 per cent, while the goose's egg which weighs 137 gm. has 14 per cent. Heinroth, and Groebbels & Mobert, among others, have collected a great many data of this kind for all varieties of bird, but their papers must be referred to for the figures. Thus the fertilised embryo starts its development on the surface of a mass of food only slightly diluted with water, and surrounded by a further and much wetter supply. This is reflected well by the work of Bellini, who found that the yolk of the hen's egg was seven times as viscous as the white at the beginning of development. (Alb. 3-4 units, yolk 28-5 units.)

A good deal of work has been done on the variability of the weights of the parts of the egg within a given species of fowl. Thus Jull found that egg weight is the least variable factor, albumen weight slightly more variable than egg weight, yolk weight considerably more variable than albumen weight, and shell weight the most variable. It would seem, therefore, as if a compensatory process takes place during egg-production, the largest yolks having the smallest whites, since the weights of the entire eggs do not vary as much as the weights of the components. On the other hand, the smaller eggs contain the highest percentage of albumen and shell and the lowest percentage of yolk. Jull also studied closely the seasonal variations, which may be quite considerable, finding that the component parts of the egg contribute in different degrees at different times of the year towards the total egg weight. The question as to which part of the egg is mainly responsible for large or small eggs is still debated, for Curtis concluded from his observations that it is the egg-white, while Atwood found many indications contrary to this. Statistical studies on the egg of the tern have been made by Rowan, Parker & Bell; Rowan, Wolff, Sulman, Pearson, Isaacs, Elderton & Tildesley; and by Watson, Watson, Pearson, Karn, Irwin & Pearson.

The division of birds into the two classes of nidicolous (those which hatch as squabs) and nidifugous (those which hatch as downy, feathered and active chicks) has been shown to extend to the composition of their eggs by several investigators. Davy found that the eggs of the nidicolous birds had thinner and more fragile shells, which took up a less proportion of the weight of the whole egg than the shells of nidifugous birds. Thus the wren's egg-shell weighs only 5 per cent, of the whole egg weight, while the hen's weighs lo per cent. Da\y's figures show very clearly that the main reservoir of solid is the yolk and the main reservoir of water is the egg-white.

Tarchanov carried the matter further, and observed that the yolks of nidicolous birds always formed a smaller proportion of the total amount of material inside the egg than in nidifugous birds. Thus, for the former class the egg-white accounts for about 78 per cent, of the egg and the yolk for 22 per cent., while, in the latter class, the egg-white accounts for about 55 per cent, and the yolk for 45. All these differences are probably related to the shorter incubationtime of the nidicolous eggs; and, as will be seen later, there are not wanting indications that the yolk of these is less tightly packed with food-material and more rich in phosphatides.

1-4. The Chemical Constitution of the Avian Egg as a whole

The composition of the egg as a whole is further considered in Table 2, where it is noticeable that the analyses of water and ash have not been significantly improved upon between 1863 (the date of the first analysis, Payen's) and the present time. The later figures for protein and fat are, however, much the more reliable. It should be observed that there is an approximately equal quantity of fat and protein at the disposal of the embryo, though the former is, of course, in the yolk, and the latter is preponderantly in the egg-white. This protein-fat equality is by no means the rule in all eggs, and, as we shall see later, the eggs of fishes depart widely from it. There appear to be only small differences between the eggs of different kinds of birds in protein content. At one time it was thought that the duck's egg was particularly rich in fat, on the authority of Commaille's analyses, but Liihrig has since then brought it into line with all the others. It does seem, however, to have a considerably higher percentage of mineral substances than the rest. The dry-weight figures merely demonstrate again the approximate equality of the protein and fat.

Before we proceed to consider the parts of the tgg in separation, the question of individual and racial differences must be taken up

Table 3. Individual differences between hen's eggs, Malcolm's figures (1902). Averages of individual hens.

Breed unknown

Italian hens (fed on maize and barley)

From one hen

From one hen

Iljin's figures (19 17).

Houdan... Orpington Plymouth Rock Rhode Island ..

itty acids



















I 52






















30- 1 2







% dry weight


Lecithin P











59- 1 6



Table 4. Race differences in hen's eggs.

Leveque & Ponscarme's figures.


% of egg d.

"y weigh



dry weight



of whole



N in



i> 111





in yolk


Andalusia ...














Bressane ...







Coucou de Rennes














Dorking ...







Faverolles ...














La Fleche ...























O ■"


J2 050 -^CTl m(N COOtD-r)"COOtDCO {S«30 I^

2 eD-*inTt<r^io<r>co co^p ^nto 'O'X) m coco in

2t; CTio o-*' COM cni^ or~ cor^ tj-od co ct> <~~ ci "SS in^<N<o ocococo co«coo oeoip >oo co ccdo"'-"'-' Of" 6" "'-I "CI o<6 66c<"

a a Tt>o com o<co into oci p~-o cocoot~4j(00 lOCT) coio i-cio -^lO oco mmtD maii-i

r^Mcoc^cooci 00)66 ctjo 6" 6<oicO" " cococococococow cococtcocococotxcoco

i- o o o o o


■53 oit^tico (r)f~-t6t6 trjLO ■ij-ti) (i> f^ t6

i-i coco i^ CI 0^0 V a><£) lo CT) t£> ■*

rt (X) C£) CO •^ CT> K


6 6 6 6 mm mm

o CO r^ o

CTi r~- (T) ■*

« en cicb m -^ ■* ■*

- CO CO c< CO CI to 05C£) oieom'^'cici o ocor^

-fi CO CO CO CO CTlCO CO CT> CO O OiCO CD p r^ f^ O CO

<6666 66 66 6« 66 6" 66«6

ot; o-*ciTi-coococi ocicoci o^n coocor^ -O 2 f^ CI coco ►"t^mcTico'^cici cocna)coCT;co

again. Malcolm maintained in 1902 that, although there were undoubtedly differences between the eggs of different breeds of hen, they did not exceed the amount of variation between individual eggs from hens of the same breed. His results are shown in Table 3. Thus, although the feeding was carefully controlled, the eggs from one hen might show a difference of i'i4gm. of fat, while between two eggs from different breeds the difference might be only 0-13 gm. His conclusions were supported in the main by Carpiaux; Leveque & Ponscarme ; von Czadek ; Iljin ; and Willard, Shaw, Hartzell & Hole; who made very long studies of a considerable number of breeds. Some of the figures obtained by von Czadek and Iljin are given in Table 5. Von Czadek studied the Sulmtal, Minorca, Orpington, Rhode Island, Faverolle, and Wyandotte races, together with an Italian and a Rhineland breed. His outside values for egg weight, for instance, were 43 gm. and 75 gm. — a considerable difference — but the former was from the Rhineland hen and the latter from the Minorca variety. The span showed great variations, thus an egg weighing 55 gm. might be a heavy Orpington or a rather light Faverolle or a medium weight Italian. The only breed which stood well out of the range of individual differences was the Minorcas which laid very heavy eggs. Certain instances have shown, however, that remarkable agreement may exist between work done on eggs of widely different breeds. Thus the classical work of Plimmer & Scott on the phosphorus metabolism of the developing chick was confirmed very strikingly by Masai & Fukutomi, who worked in Japan. Here the correspondence was almost numerical. But on the other hand there is evidence that eggs of different breeds differ not only in their gross characteristics, but also as regards more subtle properties; thus Moran has demonstrated that eggs from different breeds of hen vary very greatly in their resistance to cold, so that the viability is different, and Needham, working on the inositol metabolism of the embryo, observed differences between the embryos from Black and White Leghorn hens. Physico-chemical differences between breeds of silkworm eggs are enumerated by Pigorini.

The individual differences between eggs may be equally important. Benjamin has shown that there are numerous variable factors which modify the constitution of the egg. The amount of yolk, egg-white, and water, as well as the thickness of the shell, vary according to the season, diet, age (Riddle) and general condition of the bird in question. Nor are such comparatively slowly changing factors the only ones which bring about differences between individual eggs ; the time the egg takes to pass down the oviduct, for instance, will materially affect the amount of albumen it contains, and such variable quantities as the blood-sugar level (Riddle) and the level of cholesterinaemia in the parent animal will exercise their effects upon the resulting egg. Again, the length of time elapsing between the laying of the egg and the beginning of incubation will have a marked efTect, for a certain amount of water will evaporate from the egg-contents through the shell, and just how much does so will depend on the humidity of the surrounding atmosphere. The process of water-absorption by the yolk (Greenlee) from the white will also be affected by these conditions, so that the embryo at the initiation of its incubatory development may find a remarkably inconstant set of circumstances in its immediate environment. Moreover, a certain amount of development always takes place in the egg after fertilisation as it passes down the oviduct, so that the embryo has already gastrulated by the time that the egg is laid by the hen. It was the ignorance of these facts which led Malpighi, as we have already seen, to his erroneous conclusions, for if he had known of the phenomenon of "body-heating", as it is called by the poultryfarmer, he would not have put forward the preformation-theory, and the eighteenth century would have been spared the trouble of getting rid of that embryological phlogiston. Thus no two eggs are ever exactly the same age, and as there is reason to believe that enzymic action begins in the yolk, if not in the white, very shortly after fertilisation, this fact makes it additionally difficult to get precise figures for the constitution of the unincubated egg. Then the position of the egg in the clutch (whether first or second) in pigeons may, according to Riddle, make a difference of 9-15 per cent, in yolk weight. It may be concluded that nothing short of the greatest caution must be employed in the material which is used for chemico-embryological researches on the hen's egg. The individual hens should be marked, and the eggs produced by them should be noted, their food should be constant in composition and the breed used should be not only single in any one series of experiments, but also, if possible, genetically pure. It is very greatly to be wished that standard hens could be obtained, such as the standard rats necessary for feeding experiments, and much further work, with a proper statistical backing, is needed on the range of individual and racial variations in all the properties of eggs.

The effect of the diet of the hen on the chemical composition of the egg has been studied by various workers, notably by Terroine & Belin. Except in certain respects, it showed a remarkable fixity of composition :

Table 6.

Ordinary Corn and potato Hemp seed mixed ration almost ration

ration free from fats (fatty)

White in % of total weight ... 56-7 54-3 —

Yolk in % of total weight ... 31-3 34-0 33-2

Shell in % of total weight ... 11-4 lo-g — White

Water % 87-8 87-4 87-4

Ash% 0-49 — —


Water % 49-9 50-33 50-99

Ash% 1-48 — —

Total nitrogen % ... ... 2-67 — —

Total fatty acids % 28-4 26-6 2655

Unsaponifiable fraction % ... 1-85 — 2-08

Cholesterol % i-i8 1-58 i-ii

Lecithin P % — 0-425 0-434

Thus, although the character of the substances stored for the use of the embryo can be varied considerably, as will be seen later, the balance of them cannot. But the question is probably rather complicated, for it has been shown by Dam that by feeding hens on a ration rich in cholesterol, the cholesterol content of eggs can be increased from 501 to 615 mgm. per cent, of the wet weight or roughly by 22 per cent, of the original value. In another instance the cholesterol rose from 476 to 560 mgm. per cent. This would not be in disagreement with Terroine & Belin's figures, but it would be a very desirable thing to make a detailed study of the limits of variation of all the constituents of the egg, and to find out exactly how different in chemical composition an egg can be from the normal while retaining its hatchability. Klein regards the cholesterol output of the hen in its eggs as showing a synthesis of that substance in the parent body. Leveque & Ponscarme have stated that it was not possible to show any effect on the eggs in eleven breeds of hen by minor variations in the diet; and this was amply confirmed by Gross.

The ingenious and partially successful attempt of Riddle and Behre & Riddle to make hens preserve their own eggs by feeding them with hexamethylenetetramine, sodium benzoate, and sodium


salicylate, may here be mentioned. Starting out from this practical suggestion the work led to the discovery of a number of specific effects of substances such as quinine on egg size and yolk size. Thus Riddle & Basset found that alcohol markedly reduces yolk size in pigeons, Riddle & Anderson found that quinine reduces egg size, yolk size and albumen size but has no effect on the protein/fat ratio of the egg, while Behre & Riddle found that the diminution of albumen size under quinine bore more on the solids than on the water and involved considerable reduction of the protein.

The elaborate investigations on the egg of the tern, already mentioned, led to a significant correlation between abundance of food and size of egg, and it is certain that the size of the hen's egg is affected by its diet since the work of Atwood. There seems also to be a seasonal fluctuation, the weight of the eggs increasing from July to February and decreasing from March to June. These seasonal fluctuations appeared distinctly in Atwood's data, and explain the results of Curtis and of Fere. Rice, Nixon & Rogers and Riddle found a definite relation between the amount of food consumed and the number of eggs produced, both of these factors varying exactly with the seasonal variation in the egg size. Fluctuations of a regular kind seem even to occur each month, according to Hadley who observed such changes in egg weight and number. According to Curtis the size of the eggs increases as the laying bird matures, in the case of the hen, and Pearson has observed similar variations in the case of the sparrow.

The genetics of egg production have been studied by Pearl and Benjamin.

The relation between the egg weight and the chick weight at hatching has been studied by Halbersleben & Mussehl and by Iljin. The former workers found a quite consistent relation within one breed between the weight of the egg before incubation and the weight of the chick at hatching, the latter averaged 64 per cent, of the former. After thirty-five days of post-natal life, however, the slight advantage possessed by the chicks from the heavier eggs had altogether disappeared. They also noted that, other things being equal, chicks hatched from the more pigmented eggs (browner) weighed slightly more than those hatched from the less pigmented ones. Abnormally large and abnormally small eggs did not hatch as well as those of medium weight. Iljin collected a great many figures but his text contains no statistical analysis.


Stewart & Atwood reported that chicks hatched from pullet eggs were neither so large nor so vigorous as those hatched from the eggs of hens two or three years old. Whether there is here a direct effect on the chick of the age of the hen, or whether the effect is indirect, due to the small size of the egg, may be well questioned.

What relations exist between the chemical constitution of the egg and the percentage "hatchability" are at present obscure, owing perhaps to the comparative crudity of our estimation methods. The work of Pearl & Surface indicated definitely that differences in the hatchability of eggs are determined by or associated with innate differences in the individual hens which laid them, that these differences are probably inherited, and that variations within rather wide limits in certain environmental factors, e.g. the temperature, during incubation, are of secondary importance in determining the death or the hatching of the embryo. Hatchability of embryos would appear then to be, like fecundity, a heritable character. The experiments of Lamson & Card confirmed the conclusions of Pearl & Surface, but although some physico-chemical mechanism is undoubtedly at work, these statistical studies gave no hint as to its nature.

Dunn determined to probe further into it. In his first paper he argued that if hatchability was associated with constitutional vigour, it should show a correlation with such a value as the chick mortality in the first three weeks of post-natal life. Experimentally this was not the case, e.g. post-natal mortality remained the same, although in two instances the pre-natal mortality was on the one hand extremely high (20-39 per cent, hatchability) and on the other hand extremely low (80-100 per cent, hatchability). It therefore seemed likely that mortality before and after hatching is determined by quite different factors. The more specific influences operating in embryonic life must doubtless be looked for in the physico-chemical constitution of the unincubated egg.

Hays & Sumbardo, in a search for such influences, were able to exclude statistically fresh weight, length, diameter, specific gravity, shell thickness, outer and inner shell-membrane thickness, porosity and imbibition of water from 25 per cent, salt solution. Other factors which have been excluded are percentage of protein in the diet of the laying hen (Rosedale), percentage of yolk-pigment (Benjamin), evaporation rate of the egg (Dunn), yolk-fat percentage (Cross), egg


fat percentage (Cross), yolk-protein percentage (Cross), egg-protein percentage (Cross), egg-phosphorus percentage (Cross), chickphosphorus percentage (Cross).

It appears, however, that the constitution of the egg-proteins maybe influenced by the presence of unusual proteins in the diet of the hen, and that this may influence hatchability. Pollard & Carr have reported the results of feeding the following proteins to laying hens : wheat, rye, corn, oats; kaffir, barley, peas, soya, hemp; buckwheat, popcorn, sunflower seed.

The first group of four (all, of course, being fed alone) were very efficient for the production of normal eggs; the second group (of five) permitted the hens to lay eggs but the eggs were hardly hatchable at all, while the third group allowed of no eggs. Pollard & Carr studied the egg-proteins in all cases and obtained evidence of tryptophane deficiency in the second group, so that they concluded that a minimum tryptophane content was essential for successful development through hatching. It is unfortunate that their results were never published in full.

The effect of sex on the chemical composition of the egg has been discussed by Riddle. As is well known, in some, probably most, animals, the male produces two kinds of spermatozoa which are not equal in their prospective sex value, i.e. some which will give rise to females and some which will give rise to males. In birds, on the other hand, the dimorphism of the germs exists not in the spermatozoa but in the egg-cells. The female produces two kinds of eggs of unequal prospective sex value. Riddle found that pure wild species of doves and pigeons were ideal material for studies on sex, since very abnormal sex-ratios could easily be obtained from them, and his studies led him to the view that sex was more a matter of metabolic level or rate of protoplasmic activity than anything else. But what concerns us here are the consistent differences which he was able to demonstrate between male and female eggs.

Pigeons generically crossed, when not permitted to lay many eggs, produce only males, but when made to lay many eggs produce first only males, and eventually "under stress of overwork" only females. These facts and their proper conditions having been ascertained previously by extensive statistical investigations, the way lay open for the chemical analysis of the two sorts of pigeon's eggs. 900 analyses were made and more than 12,000 yolks weighed.



[PT. Ill

Fig. 14, taken from Riddle, gives the differences diagrammatically. A glance at it shows that the male-producing egg of the spring contains less stored material than the female-producing egg of the

Sex conbrol and known correlations in pigeons



d cf cT d* cT ^ $ ^^ *? *$ ^9 •.>' ^ •-': "■

9 10 11 12 13 14

N? 1

N° <2

N° 3

N9 4

N? 5

N° 6

N? 7

N9 8

N° 9



N0I. shows comparabive size of eggs of Alba (A) and Orienbalis (0)

Fig. 14.

autumn. The amount of water and ash present, on the other hand, diminishes, and the rise indeed is mainly to be seen in the fat and lipoid fractions and in the calorific value. Table 7 gives the figures for one individual pigeon during 191 2. The differences are not

SECT. l]



large, but they were invariably found. Another series of figures showing the rise in calorific value during the course of the year and the transition fi-om male to female eggs is given in Table 8. Here also the increase is unmistakable. Within one clutch, also, the watercontent of the second egg is lower and the calorific value higher than the first egg, which fits in very well with the fact that under normal


May 26


June 7



July 3 5

15 17 23 25


4 13 25







Nov. Dec.

Table 7. Effect of sex on pigeon's eggs.

Female Turtur orientalis x Streptopelia alba, no. 410 for 191 2.

% wet weight % dry

Analysed Weight t ^ ^ weight

or in- of Pro- Extrac- ale- Calories

cubated yolk Lipoid tein tives Ash Water sol. per egg Sex

An. 2-330 18-32 25-44 5-28 4-85 5701 72-65 7405 —

An. 2-660 17-54 25-63 5-25 2-62 54-82 72-45 8990 —

Inc. — — — — — — — — Male

Inc. — — — — — — — — Male

Inc. — — — — — — — — Male

An. 2-026 16-49 26-00 3-63 2-43 56-05 71-95 6714 —

An. 2-330 19-18 26-55 3-75 1-93 5522 72-27 7881 —

Inc. — — — — — — — — Male

Inc. — — _____ _ Male

An. 2-422 17-82 25-88 3-82 I -80 55-84 72-42 8061 —

An. 2-720 18-88 25-96 3-86 1-81 55-33 72-45 9296 —

Inc. _______ _ Male

Inc. — — — — — — — — Male

Inc. — — — — — — — — Male

Inc. — — — — — — — — Female

Inc. — — — — — — — — Female

Inc. — — — — — — — — Female

An. 2-700 21-40 — — — 55-45 73-17 9323 —

An. 2-715 21-63 — — — 55-39 73-02 9383 —

Table 8. Eggs from the same female Streptopelia risoria (1914).


Weight of yolk Energy in cals.

June 6










I -000


July 14






Aug. 30



Nov. 6












Dec. I



















conditions the first egg laid nearly always gives rise to a male and the second to a female.

In Fig. 14 the line marked "developmental energy" implies that a higher percentage of the male eggs hatch successfully than of the female eggs. The data for length of life show the same curve. The smaller eggs of both clutch and season are the eggs which give positive results in strength and vigour tests, and the larger eggs are those which are liable to display weakness. These facts are in entire accord with the higher metabolic level which Riddle associates with the small male eggs. It is interesting to note that Lawrence & Riddle found consistently higher values for total fat and total phosphorus in the blood of female fowls than in that of male fowls, from which they concluded that the metabolic differences between male and female germs persist in the adult, and all these facts are in agreement with the work of Goerttler and Baker on human and Smith on crustacean blood-fat, and of Benedict & Emmes on sex differences in basal metabolism. But for further discussion of the metabolic theory of sex, the papers of Riddle must be consulted. Interesting data on the hatchability, vigour, etc., of rotifer eggs are contained in the paper of Jennings & Lynch, but these authors made no chemical experiments.

To say, as Riddle does, that there are, as it were, two kinds of eggs in some species, one male-producing, and the other female-producing, may either be taken to mean that there are quantitative differences between them or that their constituent substances are qualitatively chemically different, or, thirdly, that the same substances in the same quantities are differently distributed spatially and temporally. As will be seen later in connection with the lipoids of mammalian egg-cells, the second view finds supporters, and some such opinion is held by Russo. Faure-Fremiet, in the course of his work on the egg of Ascaris megalocephala, to which he applied every conceivable method, examined a very large number of individual eggs in order to find whether they separated at all chemically into two types. His method was to centrifuge them separately, much as McClendon had done with the frog's egg, and then to measure in mm. the thickness of {a) the mitochondria layer, and {b) the fatty layer. Fig. 15 {a) taken from his paper shows the frequency polygon which he constructed on the basis of these results, the ordinate giving the number of eggs measured, and the abscissa the thickness of the mitochondrial layer.

SECT. l]



It is quite evident that there are not two modes on the line joining the points, i.e. that there are not two types of eggs, but only one type. Faure-Fremiet made very similar experiments, determining the glycogen content of the eggs histo-colorimetrically with iodine solutions, and there also the frequency polygon had but one mode (Fig. 15 (^)). But this second case was based on an unsatisfactory method. In the particular instance under investigation, neither mitochondria nor



12 3^5

30 20



J :

<k 5 7 8

ib) Fig. 15 glycogen happens to be an entity which varies as between the two kinds of eggs. Nevertheless, the plan of work was an interesting one, and widely extended researches with it, using accurate chemical methods, would be very desirable.

1-5. The Shell of the Avian Egg

Litde attention has been paid to the shell of the bird's egg from a physiological point of view. The relevant analyses are given in Table 9. There is some difference between the shells of different








o o


• rt £-< tJ-* 1-H

^ OCO « 


S ^


3 --^



m g .

U " 3 . -1 ^-^ u


^ ^



g a


in CO ■*

" I~» CO

6 6 6

in-*iininTj<co,coinc«oeoor^ CT)








CO mco


o o o



CO t^ -*


O .bo



r-» CO tJh o^cd o CO oj f^ tjh a>

CoiOTf<f^-^" COlOCOtj^C^ 01CT)OCnC505CT)CT)005CT)


' Um

e o u

(U O 3



• •^


C C >- D

5 s i!

b ^ 3

O C O c O U 3 « 



. . T5 i-i . . . .

. . g i) . . . .

• • Q-a • • • •

o g ffiOOOOffiffiO

P 2

«1i I

4) o O S it( t^ O « !;<

c !- 'y ,0 c«UOh



3 II

SECT. l]



birds, and it would be interesting, for example, to know why the pheasant's contains such an unusually high percentage, of phosphorus, and why the herring gull's has so high a percentage of organic substance. A certain interest attaches to the determinations of Balland on ostrich eggs, some from a tomb of the Hellenistic period and others from modern ostriches, but the differences he found were probably not very significant, as the analysis of Torrance seems to give values half-way between those of Balland. Neither Balland, Torrance nor Wicke states whether the ostrich used was the North or South African variety, a complication which might make a difference. Wicke believed that the difference in shell-composition between different kinds of birds was almost entirely dependent on their usual foods.

The microscopic structure of the shell was investigated by Nathusius in the 'sixties, and since then little has been added to his work. The shell consists of an outer layer of crystals of calcium carbonate arranged with their long axes perpendicular to the bounding surface (Fig. 16), and an inner layer composed of undifferentiated calcium carbonate (Herzog & Gonell). Kelly; Schmidt; Meigen and Osawa have found that the mineralogical form of the lime is invariably calcite, no aragonite being present in any bird's egg-shell. This has been confirmed with X-ray analysis by Mayneord. The ostrich, Emj>s europaea, is the only doubtful case, for Kelly identified its egg-shell lime as conchite, but Torrance considers it to be calcite. Prenant's review should be consulted for further details regarding this interesting biochemical problem. Only one paper exists dealing with the changes which the shell undergoes histologically during the development of the chick; it will be considered in the section on embryonic respiration, where the data we possess on the question of the

Fig. 16. a, Outer crystalline layer; b, c, d, amorphous layers; e, mamillae; /, shellmembrane.

permeability of the shell and its membranes will also be dealt with. There has been some controversy on the subject of whether the eggshell contains any elements of the secreting organ in it, like the decidua of mammals. Von Baer thought not, but the presence of cellular structures has been reported by von Hemsbach; Landois; and Blasius.

The shell-membrane has been studied chemically by Liebermann and Lindwall, who found that it consisted almost entirely of a protein, the percentage composition of which agreed very closely with keratin (Table loa). Krukenberg, alone, on the ground of its reactions, held it to be a mucin. This ovokeratin, which contains four times as much sulphur as the albumen of the egg-white, was found by Morner to include 7 per cent, of cystine, but there are reasons for supposing that this figure is much too low. Nothing is known of the part played by ovokeratin in embryonic metabolism, but, in view of the fact that calcium is transported from the shell to the embryo during the period of ossification of embryonic cartilage, it is not impossible that the sulphur or the cystine of ovokeratin may be made use of in a similar manner to meet the need for sulphur and cystine for the feathers. This will be discussed later under the head of sulphur metabolism (Section I2'7). Morner considered that sulphur must exist in the ovokeratin in other forms besides cystine, for that amino-acid would not account for more than a third of what he found was there. The amino-acid analyses of ovokeratin are placed in Table 1 1 ; they are due to Abderhalden & Ebstein, and to Plimmer & Rosedale, the former by isolation and the latter by the van Slyke nitrogen distribution method. The arginine figure is rather high.

The strength of the shell is clearly an important biological factor : according to Romanov its average thickness is 0*3 11 mm. giving a breaking-strength of 4-46 kilos. The relation between shell-thickness and breaking-strength is a straight line.

The physiological properties of the shell of the bird's egg have been very insufficiently studied. In the last century there was a general impression that the shell possessed a differential permeability and that, while water and other liquids would readily go through the egg-shell and its membranes from outside in, they would not easily pass from the inside to the outside. It is difficult to find how this idea originated; thus Ranke in 1872 attributed it to the younger Meckel, and Ranke's own statement was subsequently copied down


siuaQ ^ ujio^

aioonwoAO :-i3ii3Z



V V t-, S-.

lOM lO t^ t^ CT) J^

'-' <-o 6 6 6 6 ,'-'

(moo) NIHSVD :u3piEqjapqY

CCTlOLOi-iiNar^iNuOO lOCC cc


s3uof jj sujoqso puB jqosjinBjg :§

qwo.wusABsq 'auioqsQ


NmaxiAOAO M o ^^ o coco m ci

(3uiui3jb) uipaH pue ;-|2e»«cooi6w jajunjj 3? uspjBqjapqv fn "" "




X3 D.

-3S E £^'




aiDV oiNiaivsAq

aiDv aiNiaiYioad


4. (psjBinSEOo) NHIMiiaiV -OAO : (auusXD) jaujoj/v[ puB pjoi\ib§ buaunoSn]^

" I o CO o « I 6 6 -^

o c< , o

O O CO ^J* c( cj ,1, I CO

CO ■^ CO

c< 6 CO

CI «


CJ 1^ lO

ClO CD 'f





uO CD C^

1 ?-^l 1



' 6 t- ' '


^^(pasTiiBjsiUo) NHwriaiv v q -OAO :u3piBqj3pqy puB g «  jSajj ^ uap]Bqj3pqv 2; =0

t^ CI tJ* — CO

(p3sn -jeisAjo) NraiMnaivoAO luaqjiQ Jg aujoqsQ puB q;jOA\u3AB3q 7g sauof 'sujoqsQ ijqoapnsjg 3g qUOA\U3AB3'7 '3ujoqso

M 01 O CO lO CI CD

>- " O ■* CO " r^ C) r^ CO CI ^

« t}< CO " " ,1^


: : : :'S o cs : : : : : .S E u S 2 o

V V V „,'rt.yS c ^c3-i33o«CI.3-r;SHMc/,b0vic3 >-"e -^ -i^ n (i:




I I I I I I I I ^1 I I I



O 7J"Ocp o

(pajEinSEOo) NHwnaivoAO

pjBUlIIBQ jj buaunoSnjj


3U}3{j 3g UBUideqQ



SjaqsjY ^ auaAsq

c T « 

' rsi

r (3>i^is


UBA) aioonivoAO



(35lXlS UBA)















-oaj aiiH.\\-003

(ajjXis UBA) SNiaiOHd-HiOA

o o o o o , o

CO O OtX> O 7**

eoTj^i. 6 " 6

or-6 "

CO a> ;*•


CO 'f CT)









-a c












^ 1)

t^ CO


Ph sh

C< Tf


T ^



t^ !N



o e<





'in O

01 r-^


M t« 

c< o

« « 


r^ c<




-]pqs) NI1VH3M0A0

(aupsAo) j3ujoj\i puB uiajsqg

^ uapjBqiapqy


?^ 9

UBA) SNiaioHd-ooa

aaxiw :s3puMoq

^ jauiuijitj

+ o + 11 CO -^coo 1 1

' ' CO o


« o Th« 

LO •+■ 01U3


.s.s H.S s->-g i

>-nil5 j^ O W Dh_3




LNaiMiiaivoAO I i I I I I I I I



I 1 I

I I 1 ?. I I 1 I


(suusXo) uBAT^ng puB (auBqdojdXai)

1 I I I I I l|l I I I I r I

NHwnaivoAO ^



e-6 i u i •: I 1 1 1 1 1 1 I I I

!r> CO

o ^f


I r.

I I I 1 1^ I I I I




aaxiw :3duiax ■„ I 28 uapieqjapqv


-OAO :s3UJ30H

jg jauiBSuapv

C U OJ OJ o 3 n m


aaxiiM :nfpu3S

to o» a ■* 1^


-IVXOO UH33I -U3AB3T 'aujoqso











osine ... tine ... tidine


'.3 "S

c S


■ptophane identified £ al di-amin


S 6 — t« 








wrongly by Schafer. However, Thunberg in 1902 conclusively demonstrated the error of the belief, and showed experimentally that water would pass through the membranes equally well both ways, though he found that of the two the inner one was the less permeable. In the case of water birds, there is evidence that the shell is absolutely impermeable to water; Loisel, for instance, found that the eggs of the grebe, Podiceps cristatus, and the duck, Anas domesticus, when placed in distilled water absorbed not a trace of it, and gave out no chloride to it over a period of many hours. The eggs of the ordinary hen, on the other hand, increased considerably in weight and allowed some chloride to pass out into the water. At the same time, Loisel found that hen's eggs develop quite normally, at any rate up to the seventh day, even if they are lying in water. This experiment of Loisel's was confirmed by Lippincott & de Puy, who succeeded in hatching chicks from eggs incubated while lying in | in. of distilled water. The differences between these eggs and the controls suggested that the eggs lying in the water not only failed to lose as much water through evaporation as eggs incubated under ordinary conditions, because of the limitation of the evaporating surface, but actually absorbed some water. Trials with rhodamine red and methylene blue demonstrated penetration by these dyes, extending in the former case to vital staining of the embryo. It is known (Rizzo) that the avian egg-shell has many pores (o-86 to 1-44, average 1-23 per sq. mm.).

As regards gases, the only paper is that of Hiifner. Hiifner placed small pieces of egg-shell with their membranes in a diffusiometer, and measured the rate at which gases passed through the obstacle. He found that oxygen diffused through with most difficulty, then nitrogen, then carbon dioxide, and, finally, hydrogen most easily. It may be significant that carbon dioxide would thus appear to be able to escape somewhat quicker than oxygen can enter. But under normal atmospheric pressure the amount diffusing through the whole egg-shell (goose) per second was 2-115 c.c. of oxygen and 0-503 c.c. of carbon dioxide. The diffusion velocity was always proportional to the partial pressure of the gases, and the removal of the inner membrane made no difference at all, suggesting that the principal barrier was the amorphous calcium carbonate layer. It would be very desirable to repeat these observations with more modern methods, and on a greater variety of eggs.


1-6. The Avian Egg-white

The white of the egg is divisible into three portions which have been studied separately by Romanov. The outermost and thinnest layer makes up 39-8 % of the whole and has 1 1'6 % dry solid. The middle layer accounts for 57-2 % and has 12-4% dry solid, and the innermost, thickest, layer is only 3 % and has i4'5 % dry solid*. The chalazae have only once been analysed separately (Liebermann), when the elementary composition of their protein was ascertained. Table 1 2 summarises the results of the investigators who have made general analyses of the white. It is a very watery solution of protein, containing only the most negligible traces of fats and lipoids, but a great many water-soluble substances such as carbohydrate in various forms, protein breakdown-products, choline, inositol, etc. Natural egg-white, according to Rakusin & Flieher, is a saturated solution of ovoalbumen (15-35 P^^ cent.). The water-content does not vary much, but Tarchanov's analyses go to show that the smaller eggs with short incubation-time are wetter than the others. The proteins of the egg-white are believed to be variable in number in the eggs of different birds. In that of the hen, four are known, in that of the crow three, and in that of the dove one only. The egg-white of the hen's egg contains two albumens, ovoalbumen and conalbumen, and two glucoproteins, ovomucoid and ovomucin.

It was at one time thought that there was a fifth, ovoglobulin. Dillner studied it in 1885, and estimated that it made up 0-67 per cent, of the egg-white and 6-4 per cent, of the total protein, but Osborne & Campbell showed that it was simply a mixture of the others in different proportions. This had already been made probable by the results obtained by Corin & Berard, who were able to separate the ovoglobulin into two or three constituent proteins having several different coagulation temperatures (57*5°, 67°, 72°, 76° and 82° C.) and other special characteristics.

Hofmeister was the first to prepare crystalline ovoalbumen, and he published several papers on it. Other workers confirmed his discovery, such as Gabriel ; Harnack ; and Bondzynski & Zoja, but Hopkins & Pincus showed that the albumen so crystallised only accounted for half the protein present in the egg-white. Part of the missing protein was found by Osborne & Campbell to be in the

A large number of concentric rings can be seen in egg-white coagulated in situ, according to Remotti.



[PT. Ill

2 U > B

c ^

he- O 1) > •O O g :3 §

o M

W a o

tj> CO

Ph O ■ - (u D c u 2 9 o



O ^Z,C^ CD "

, e< ' > . ,„ n

3 " 5-5 S cng ^

>-j >j= ^ to.c e £

Ki cij u .^ bc:^ tS i=!






'JD O — r^ I CO r- I I

o I •* « ' CO



— . O CO LO r^ f^ CO xi LOCO 6 6 6 6 6

o in

o o


^D^D I 4^ f^


? 2 u A ?^ =« 



r^tr> oco(NCOco, "^.cjoc*

0(C)c<oooo 7" -^ p 6666666 6 66

" o 6 6


" oco x^ CO ■* CO




!N CO 0» CO CJ


- - « 


cr> 0!n

CO 0)0 I^ ot

000 CO t^UD

M o -' o 6


rS bn w

C bo Q <fi u OS

(0 D


t, -2 c

3 D W



C3 Q^ ^

S"! K i-i c3 o « 



< 3 "=o o
















r- oi


If I I I I I I I I I I i I I I I i I I ?

OCO" 000000000 0000000 CI N

cocTi^ «a>-<o~ciLOcoyf oito c~,co c~. o o c> o

cococo cocococoo oico coco cocooococococococo

o ^ ^

jn v X . t„ to

u '^


Table 13. Distribution of proteins in avian egg-white




Passeres (singing birds) Turdus iliacus Turdus piliaris ... Anthus pratensis ... Anthus cervinus ... Garrulm infaustus Corvus cornix Corvus nionedula ...

Zygodactyli (2-toed) Dryocopus martins

Accipitres (birds of prey) Strix aluco Falco gyrfalco Falco peregrinus ... Falco aesalon Falco tinnunculus... Astur palumbariiis Buteo lagopus Pandion haliaetiiis

Pullastrae (doves) Columba livia

Gallinae (fowls) Gallus domesticus Meleagris gallopavo

Grallae (marsh birds) Charadrius apricarius Haematopus ostralegus Tringa alpina Phalaropus hyperboreus Totaniis glareola . . . Actitis hypoleiicos Limosa lapponica Numenius arquatiis Numenius phaeopus Fulica atra

Lamellirostres (ducks) Anser segetum Fuligula marila ... Fuligula fuligula Oedemia fusca Clangida glaucion Somateria mollissima Mergus senator . . .

% wet weight


% of total protein in eg5-white






1-78 1-85

2-33 2-09 1-56

1-73 1-58



1-53 1-26

1-55 1-28


2-II 1-25


I "49

i-i6 1-40 1-05 1-03 1-32 1-31 1-26 1-71 1-54

1-57 2-o6

1-45 1-40 1-56

2-00 1-67

OVO albumen 80

80 80

Investigator and date

Osborne & Campbell

(1909), r^

Komori (1920) Needham (1927) Morner (19 12)

— Morner (1912)

12-50 18-25



Table 13 [cont.].

% wet weight


% of total protein in egg-white


ovo- ovo- ovomucoid mucin albumen

Investigator and date

Steganopodes (pelicans) Phalawcrocorax carbo Phalarocrocorax graculus ...

0-2I 0-46

2-o8 — —

Mdrner (191 2)

Longipennes (swallows) Larus cantis Lesiris crepidata ... Sterna macrura ... Sterna hirundo

1-76 1-28 1-36


15-00 — —


Pygopodes (divers)

Podiceps cristatus Colymbus arcticus

2-04 1-88

— — —


form of the two glucoproteins and the other albumen, while part of it was accounted for by the fact that the yield of the crystallising process is not great. Ovomucoid was originally discovered by Neumeister, who called it "pseudopeptone", and first studied by Salkovski and Zanetti.

The investigations of Osborne & Campbell, whose memoir is the best on this subject, give no very definite indication of the proportions in which these proteins make up the protein fraction of egg-white, but they put ovoalbumen at about 80 per cent., and ovomucin at about 7 (see Table 13). Later, Komori estimated that ovomucoid accounted for about 10-5 per cent, of the proteins, and in 1927 I obtained a figure of 7-6 per cent, for the same constituent. Morner, in his extensive study of ovomucoid in numbers of birds' eggs, obtained results from which higher figures emerge on calculation, namely, from 10 to 20 per cent. The only exceptions were the pelicans, which seemed to have very little ovomucoid. The most probable relationship between the proteins is as follows: ovoalbumen 75, ovomucoid 15, ovomucin 7 and conalbumen 3 per cent., but these values are only very approximate, and further work on this point is much to be desired. Leaving out ovomucin, Wu & Ling found that the proportions were as follows (for Gallus domesticus) : ovoalbumen 78-3, ovomucoid 12-3 and conalbumen 9-4 per cent., or 1-34, 0-21 1 and o-i6i gm. per cent, respectively. Certain Russian workers (Worms and Panormov) have described two proteins, anatin and anatidin, in the egg-white of the duck's egg, and three, corvin,


corvinin and corvinidin, in the egg of the crow. It is not certain, however, to which of the well-known proteins of the hen's egg-white these others correspond. Judging from the percentage composition tables in Table lo a, the columbin of the dove's egg corresponds to hen ovoalbumen and to duck anatinin, while duck anatin corresponds to hen ovomucin, but in the absence of definite information the question must be regarded as unsettled, and would repay further investigation.

The minimal molecular weight of ovoalbumen, according to Cohn, Hendry & Prentiss, is 33,800 (Marrack & Hewitt suggest 43,000), and its percentage composition is seen in Table 10 a; the best analyses are probably those of Osborne & Campbell, who give an account of its general properties. It has been further analysed by several workers who have determined the proportions of its constituent amino-acids, and whose results are seen in Table 11. The hydrolyses of Osborne, Jones & Leavenworth; Osborne & Gilbert, and of Abderhalden & Pregl were all done by acid, but those of Hugounenq & Morel and Skraup & Hummelberger were alkaline, the former using baryta. The figures agree accordingly, and all that can be said of them is that for purposes of calculation the amounts of amino-acids must be taken as minimum in each case. Attention may also be drawn to the less complete analyses of Chapman & Petrie and Hugounenq & Galimard and to the analysis of mixed egg-white proteins by Plimmer & Rosedale, using the van Slyke technique. The large amounts of hexone 4Dases found by them contrast with those found by the remaining workers, using direct isolation, and if this is not due simply to difficulties of technique it may lead us to expect a high content of hexone bases in conalbumen and ovomucin when they come to be analysed.

In Table 10 a the results obtained by Gupta on the hydrolysis products of ovoalbumen are given (see also Rudd). It is noticeable in them, as in the analyses of ovoalbumen itself, that they contain a high proportion of sulphur, though not so much as ovomucoid. The spontaneous evolution of hydrogen sulphide by egg-white on standing has long been known, and was made the subject of a paper in 1893 by Rubner, Niemann & Stagnita, who found that 100 gm. of egg-white gave off when boiled with water 10-7 mgm. of HgS. Hausmann later decided that its source must be some labile sulphydryl grouping in the ovoalbumen molecule. In 1922 Harris


observed that raw egg-white was quite non-reactive towards the nitroprusside test for sulphydryl groups, but that immediately upon coagulation by heat it became vividly reactive, and gave an intense purple colour. This change only took place in conditions where denaturation of the protein was involved, and Harris suggested that this treatment might unmask a thiopeptide linkage or some similar arrangement which by hydrolysis or keto-enol transformation would give rise to an active sulphydryl group in the resulting metaprotein molecule. Later, Harris found that only 14 per cent, of the sulphur in ovoalbumen could be accounted for as cystine, so that some unknown sulphur compound must be present in considerable quantity, and an exactly similar finding was later reported by Osato for the egg-membrane protein of the herring. The cystine recoverable from serum albumen, on the other hand, accounted for 86 per cent, of the sulphur there. The possibilities of these facts with relation to the metabolism of the embryo have not yet been explored. Philothion, according to de Rey-Pailhade, exists in the egg-white of the hen but not in that of the duck.

The principal investigation of ovomucoid is that of Morner. He had previously discovered that percaglobulin, a protein extracted from the unripe ovarial fluid of the perch {Perca fluviatilis) would precipitate ovomucoid from its solution. With this reagent he made an examination of a wide variety of birds' eggs, in order to study the distribution of ovomucoid. By direct estimation he found it to be present in all the eggs he studied, but it seemed to exist in two sharply distinguished forms, one which would give a precipitate with percaglobulin, and another which would not. Thus the hen [Gallus domesticus) with i -46 per cent, of ovomucoid gave a highly positive reaction, but the hawk {Astur palumbarius) with i -45 per cent, gave none at all. Preparations of ovomucoid from the two varieties of egg-white (see Table 10 a) did not show up the existence of two obviously different chemical individuals, and it was concluded that the preparations were in each case mixed with a small amount of the other substance. Moreover, of the egg-whites which gave a positive reaction with percaglobulin, some contained ovomucoid precipitable with Esbach's reagent (e.g. Clangula glaucion and Somateria mollissima) and others contained an ovomucoid which could not be so precipitated (e.g. Gallus domesticus and Podiceps cristatus). It is quite uncertain how many of the effects observed by Morner are due to



[PT. Ill

physical and colloidal rather than to chemical differences, and the whole question should be reinvestigated. There seemed to be no special significance in the distribution of the ovomucoid which was precipitable by percaglobulin ; thus it was present in the accipitres, grallae, lamellirostres, longipennes and pygopodes, but not in the passeres, zygodactyli, pullastrae and steganopodes. As for the fowls, it was present in the eggs of the hen and pheasant, but not in those of the guinea-fowl. Morner was inclined to agree with Milesi's view that ovomucoid did not exist as such in the natural egg-white at all.

Table 14. Variations in properties of avian egg-white.

Coagulation point

Consistency and

of the

egg-white m

colour of coagulum

degrees Fahrenheit



Hard white




Pretty firm, bluish



Soft, white, translucent .


Missel thrush

Soft, transparent



Firm white



Soft white



Pretty firm, greenish, transparent


Golden-crested wren

Soft, bluish, semi-trans parent


As has already been observed, Sir Thomas Browne was one of the first to note that the coagulated egg-white of the gull's egg was quite different in consistency and translucency from that of the hen's egg. In 1863 Davy collected some data on these points, which are shown in Table 14, and Tarchanov devoted much time to the question in the 'eighties of the last century. He found that the whites of many kinds of eggs would not coagulate in the ordinary way on boiling, but either remained liquid and transparent or else set to a watery translucent jelly. This he called " tataeiweiss ", and as he went on to examine the distribution of this property he found that it was associated with the hatching quality of the bird in question. Thus all nidifugous birds, whose chicks are born fully feathered ("downy") and soon leave the nest, had eggs with ordinary egg-white, but all nidicolous ones, whose chicks are hatched as "squabs" or naked and weak, and have some development yet to complete, had eggs with uncoagulable or transparent egg-white. Thus the sand-martin, linnet, finch, thrush, canary, crow, dove, rook, nightingale, robin, starling (roughly passeres and pullastrae), all had tataeiweiss] while the hen,


duck, goose, guinea-fowl, partridge and corncrake had ordinary white. This classification agreed roughly with Davy's high and low coagulation points for the egg-white, and corresponded on the whole to Morner's two classes, the former having ovomucoid not precipitable with percaglobulin and the latter having the precipitable variety, but to this there were some exceptions; thus the plover's ovomucoid could be so precipitated, but its egg-white was tata and it yet produced fully-feathered chicks. It was, however, the only exception to Tarchanov's generalisation, (It should be explained that the word tata was derived from the name of Tarchanov's small daughter.) Tarchanov found that tata egg-white was about 3 per cent, richer in water than the other kind, a conclusion which Morner's later analyses did not confirm. He also said that it was alkaline to litmus, but became less so as the tata eggs developed. This agrees with the later classical work of Aggazzotti on the reaction of the eggwhite of the hen's egg during its development. Tarchanov reported that tata egg-white could be made to coagulate at ordinary temperatures by the addition of a little potassium sulphate, and that it would itself coagulate if the temperature was raised well above the boilingpoint of water. It was, he said, more easily digested by enzymes, it putrefied more readily, and during development it changed into a form resembling ordinary egg-white. He made some studies on its secretion by the oviduct of these birds, and was the first to perform the experim.ent of putting a ball (in his case a lump of amber) at the top of the oviduct and seeing it emerge at the bottom with layers of egg-white and a shell secreted around it. The change during development from tata to ordinary egg-white Tarchanov found he could imitate by bubbling carbon dioxide through the original white, after which it would coagulate in the usual way. On the other hand, he found that if he soaked normal hen's eggs in a 10 per cent, solution of alkali the white took on the properties of tata egg-white, and became just like the glassy material in the sand-martin's egg. He suggested some relation between these phenomena and the alkalialbuminate of Lieberkiihn, but did little to determine its chemical relationships. He was unable to get any development in the case of hen's eggs soaked in alkali.

In 1 89 1 Zoth took up the whole question of tataeiweiss once more. He was led to do so on account of some researches which he had been making on the effect produced on serum-clotting by various


concentrations of alkali, and which showed that the clot could vary very greatly in its properties, from opacity to almost perfect transparency, for instance. Tarchanov had decided that the transparent coagulum of the nidicolous egg-whites was not to be identified with that produced by sodium or potassium albuminate, but Zoth succeeded in showing that the differences were not sufficient to distinguish them. Zoth fully confirmed Tarchanov's finding that ordinary egg-white could be made to pass over into nidicolous egg-white by treating it with i o per cent, potash in the cold for ten days, and was able to explain all the differences between tataeiweiss and alkali ovoalbuminate as due to variations in the amount of alkali present, or rather the amount of cation as compared to anion. It is most unfortunate that we have no detailed ash analyses of the egg-whites of nidicolous birds, for, as will later be seen, the egg-white of the hen has rather more total anion than total cation, and this relationship might be expected to be even more strongly marked in the case of nidicolous egg-white, perhaps^ indeed, as much as to counterbalance the excess of cation over anion in the yolk. There can be no doubt, however, that the egg-whites of nidicolous birds are relatively richer in alkali than are those of others, and it is this, combined with their different water and total ash content, which causes the albumen to coagulate differently from those of others. Thus, if 5 c.c. of filtered egg-white from a fresh hen's egg be put in each of three small Erlenmeyer flasks, 2 c.c. of water added to A, 2 c.c. of 0-89 per cent. KOH to B, and 2 c.c. of a mixture of equal parts 0-89 per cent. KOH and 0-66 per cent. NaCl to C, the coagulum in A will be the usual white, thick, solid and opaque mass, while the other two will be transparent like tataeiweiss, slightly opalescent, more or less liquid, and Cmore opalescent than B. It would be interesting to reinvestigate the whole question anew in the light of recent knowledge and technique.

Another curious effect was noted by Melsens and Gautier. Melsens found that, if a stream of carbon dioxide, hydrogen, nitrogen or oxygen, was passed through dilute egg-white, or if it was shaken violently, a precipitate of fibrous membranous shreds was formed. Gautier observed that about i -5 per cent, of the protein was thus changed; he filtered it off and determined its elementary composition, which showed nothing remarkable. He concluded that a protein which he called " ovofibrinogen " existed in the egg-white, and even suggested that an " ovo thrombin " was present to turn it into "ovo


fibrin". He apparently thought that the ovofibrin was incorporated without change into the substance of the embryo. The subject has not received any attention since the time of Gautier, but it is probable that this phenomenon is explained by the work of Young; Dreyer & Hanssen and others, on the high instability of protein solutions.

Peptones were reported by Reichert to exist in fresh egg-white.

Wu & Ling have recently studied the coagulation of ovoalbumen by strong mechanical agitation. The fact that conalbumen is not coagulable by such means gave them a method of estimating it in egg-white. Thus they obtained the following figures for Gallus domesticus egg-white:

Nitrogen Total N (ovoalbumen + conalbumen 4- ovomucoid) 1-71 gm. %

After shaking (conalbumen + ovomucoid) ... 0-372 ,,

After shaking and heating (ovomucoid) ... 0-211 ,,

Coagulation of ovoalbumen by shaking was not separable into two stages (denaturation and agglutination) like that by heat or alcohol. The isoelectric point of the protein was the most favourable for shaking coagulation (/?H 4-8) and the Q^^q of the reaction was i-g. Piettre has published a method for separating the proteins which involves the use of acetone.

The relationships between the avian egg-white proteins have been the subject of some interesting immunological work. The earliest investigators who crystallised ovoalbumen found that perfect freshness was necessary, for at room temperature the crystallisable protein gradually turns into a non-crystallisable one. Bidault & Blaignan found that this process could be arrested by placing the ^gg at 0°. Sorensen & Hoyrup suggested that the protein formed was conalbumen and wished to look upon the latter as a product of ovoalbumen. Hektoen & Cole, however, first showed that though ovoalbumen was distinct from the serum albumen of the hen immunologically, conalbumen was not, and then went on to demonstrate that during the loss of crystallisable ovoalbumen which takes place as the egg ages, there was no corresponding increase in conalbumen. We must therefore look upon the latter as probably identical with the serum albumen of the adult: and perhaps only present in the ^gg owing to the inefficiency of the oviduct.

The analyses of ovomucoid and Eichholz's ovomucin, as well as the fragmentary one of conalbumen, will be found in Tables i o a and 1 1 . Willanen found that ovomucoid was much more susceptible



to hydrolysis by pepsin than by trypsin (see later under enzymes and antitrypsin). For its properties see the papers of Morner and Neumann. Both the glucoproteins have twice as much sulphur as ovoalbumen. Their carbohydrate content has been the subject of a great amount of discussion and experimental work. Berzelius was the first to draw attention to certain similarities between the breakdownproducts of sugars and proteins when acted upon by boiling acids. In 1876 Schiitzenberger asserted that the ovoalbumen molecule contained a carbohydrate group, basing his views on positive results with Trommer's test after total hydrolysis. In later years a number of workers supported the view that the carbohydrate was glucose, using in different cases methods of varying reliability, e.g. Krukenberg in 1885, Hofmeister and Kravkov in 1897, and Blumenthal ; Blumenthal & Mayer and Mayer in 1898 and 1899. Spencer and Morner, however, failed to get any evidence of a carbohydrate group after hydrolysis, and reported their negative results in 1898. Weiss, about the same time, thought he could identify a methyl pentose among the hydrolysis products, but he was never confirmed. Seemann was the first to announce that the carbohydrate was glucosamine, and his discovery was quickly confirmed by Frankel and Langstein. These later workers began to attempt quantitative estimation of the sugar, and their figures are given in Table 15. Pavy, using the then recently discovered osazone technique, made a study of a variety of proteins, and showed, as might be expected, that the yield from ovoalbumen was always greatly less than from ovomucoid. Eichholz obtained glucosazone from ovoalbumen, ovom.ucoid and ovomucin, but not from either serum albumen or casein. On the whole, it is most likely that ovoalbumen contains extremely little glucosamine, and the figures recorded in the literature for this are probably due to contamination with ovomucin. This is the view of Osborne, Jones & Leavenworth, for neither they nor Osborne & Campbell obtained any glucosamine from their very carefully purified ovoalbumen. Komori has prepared from ovomucoid, and Frankel & Jellinek from ovoalbumen, polysaccharidelike substances which they regard as the prosthetic group containing all the glucose. Following this up Levene & Mori have prepared a trisaccharide containing glucosamine and mannose from egg-white. Ovoalbumen contains 0-26% of this substance, coagulated egg-white 1*9%, and ovomucoid 5*1%. According to Levene & Rothen the

SECT. l]



molecule consists of four trisaccharides each containing one molecule of glucosamine and two of mannose.

Table 15. Ovoalbumen and ovomucoid glucosamine content.

Hen {Callus domesticus)

Hofmeister (1898)

Seemann (1898) ...

Langstein (1900)

Willanen (1906) ...

Pavy (1907)

Samuely (1911) ...

Neuberg & Schewket (1912) ...

Zeller (1913)

Needham (1927)

Abderhalden, Bergell & Dorpinghaus

Neuberg (i 901) ...

Blumenthal & Mayer (1900) ...

Izumi (1924)

Tillmans & Philippi (1929) Turtle ( Thalassochelys corticata)

Takahashi (1929)

Bywaters (1909) Seromucoid Pavy (1899) Ovomucoid







3-4 3-7



15-0 34-9

I9-5-22-3 21-7 29-4 24-0 33-7 II-5

7-4 6-2

Glucose % by hydrolysis with

Osseomucoid ... Tendon mucoid Ovoalbumen ... Serum globulin

10 % KOH

I2-0 12-3


33 2-6 2-8

5 % HCl 15-9


lo-B II-6



The free carbohydrate of the egg-white has also received a great amount of attention from an early date. In 1846 Winckler isolated a quantity of a sugar (4 grains) from the egg-white of the hen's egg, and identified it as lactose. Physiologists", he said, "will be able to tell me whether this is of importance for the embryo or whether it was some abnormality." The observation has not since been repeated, and it is in the highest degree unlikely that any lactose was ever in an egg, unless the diet of the hen was a very unusual one. Budge and Aldridge soon were at work on this subject, the former concluding that the carbohydrate was glucose, but suggesting that it might form a disaccharide during development, and the latter making no concession to Winckler. The presence of glucose was afterwards abundantly established by the work of Barreswil ; Lehmann ; Meissner ; Salkovski ; and Pavy. Later many quantitative estimations were made, and these are collected together in Table 16. The older figures for free carbohydrate may be regarded as fairly








• ?^



































5 o


- 1^ bo

iS S


- bD ■ bo



^ o bO

1 c


O 8




in N I CO f^

coco I I I I I

^ I U-)Co

CO Lo « >ti

CO I CO O ■-"

05 CO t\ o)

►1 ' M Cr^ C4

CO OsCTi I «- O CO I I^ CT) IT) CO "J^ 04

in o



> rt g

G u O O










bo bo

y bO I) w

o o o o o o o

O f< Crv o, JN, ^ Cl C< i~i -"i -^ --i IN >^


o <o Cr^l^

s « 

s 5



« S P

^ S; i>H '^ .^3 ^ ^ ti. g '^ tts tt, O O CO

" -^ "^ "^ « s . ^

O G ^

Oh ? £



i~ iS it- V









C „ O u fi „ o - O 3 W "

"bO bo



00000^ six IN O O5 tx"5

•S'S § 3

o o ,0 ,0 ,0 00 o

U "^ -« s ^

? 3 3 eg

« ,a,» K c a

S S ^55 ^ CJJ CJ Cj

Ci u 55 1^ [^ tc, fcc, -^ cq Q,





>*i •-<


S2 3 •



■2 ?<> :



C 3i






•^ 0^



• 2


J2 a

^ a -<a.




§ ■'2






^ ^

■« a « 








Chara Haem Tring


trustworthy, but not for combined sugar, in view of the demonstration of Holden that all copper-reducing methods are seriously interfered with by the presence of amino-acids and protein breakdown products. No method at present in use gives satisfactory results in those conditions, but the most reliable is that of Hagedorn & Jensen. No estimations of total carbohydrate in egg-white alone at present exist, but there is a single figure for glycogen due to Sakuragi. Morner found no evidence of fructose, pentoses or maltose.

A curious phenomenon : the fluorescence of egg-white has been reported by van Waegeningh & Heesterman, but it only occurs if the egg is not perfectly fresh, and is therefore probably not physiological.

I '7. The Avian Yolk

The vitelline membrane was investigated by Liebermann in il who found that it consisted almost exclusively of keratin. This he purified, and, having freed it from ash, made an elementary analysis of it, which is shown in Table loa. Some experiments which demonstrate the peculiarities of the vitelline membrane have been devised by Osborne & Kincaid. They found that a fresh yolk floated into distilled water, o-g per cent. NaCl solution, or glycerol, behaved exactly like a red blood corpuscle in that it swelled up and burst in the former, and shrank to a corrugated globe in the latter, while in the isotonic salt solution it remained unchanged. But with other treatment, nothing took place which corresponded to haemolysis. If the yolk was put into lo per cent. NaCl solution, it did not shrink, as had been expected, but swelled up, owing to the penetration of the saline and the consequent osmotic pressure due to the dissolving of the vitellin in the saline. This showed at once the scleroprotein nature of the membrane and its impermeability to vitellin even when in solution. The membrane is also impermeable to phosphatides and fats dissolved in ether, for if a yolk is put into ether it sinks and swells, so that the upper pole is distended by an accumulation of deeply pigmented ether. But until the yolk bursts, as it eventually does, not a trace of pigment or other substance passes out into the ether, and the same results were found with chloroform and carbon disulphide. In alcohol, on the other hand, there is no swelling, for the alcoholic solution of phosphatides and other bodies can pass out through the keratin membrane. It would be very interesting to make a more extended study of the osmotic properties of the vitelline membrane (see in this connection Section 5*6).

The yolk of the egg was investigated earlier in the modern period than the white. We may pass directly, excluding the curious analysis of the eggs of Struthio casuarius by Holger in 1822, to the papers of Gobley, which appeared from 1846 to 1850, and which, with those of Valenciennes & Fremy from 1854 to 1856, still remain models of embryo-chemical work. "John, a German chemist," said Gobley, "appears to have been the first to occupy himself with serious researches on the yolk of the egg. The chemists who preceded him considered it as made up only of water, albumen, oil, gelatine, and colouring matter; such was the opinion of Macquer, Fourcroy, and Thomson. John concluded from his experiments, which he published in 181 1, that the yolk was composed of water, a sweet yellow oil, traces of a free acid which he thought was phosphoric acid, and a small amount of a brownish red substance, soluble in ether and alcohol. Besides these he found gelatine, sulphur, and a great deal of a modified albuminous substance." Gobley referred also to the work of Prout, of Chevreul, of Berzelius and of Lecanu, who discovered the presence of cholesterol in yolk in 1829.

Gobley himself found in the yolk nearly all the substances which we now know to be there. His own list of them ran as follows :

1 . Water.

2. An albuminous matter, "vitellin",

3. Olein.

4. Margarin.

5. Cholesterin.

6. Margaric acid.

7. Oleic acid.

8. Phosphoglycerilic acid.

9. Lactic acid.

10. Salts such as chloride of sodium, chloride of potassium, chlorhydrate

of ammonium, sulphate of potash, phosphate of lime, and phosphate of magnesia.

11. A yellow colouring matter and a red colouring matter.

12. An organic substance containing nitrogen, but which is not al buminous.

Most of the constituents of egg-yolk may be recognised under this






V se

9 3't=

Z bo


■^ ^



V a

►S '-°

-G „

i^g^l J35 2 5 w

C^ ^^ O Ph ^ Ph C/D 1-1

0~i c " — o

s ^


1 •: 1 1 1 1 1


u bo


a> CO lo CO oi

1 1 M 1^ C4 O

1 ' io 'i' LO(-b io



O S"! S?

L4 U U

c 3r,

CO CO coco CO 6 6 « 6 6

CO O CO CO -^ ^ C< C< C< C4 W CO

O O "

■*Oc<r^coO'-'OOooo'-' M Oco lO" 01 coc<co cox> a in CO ■^ ^cb c) M M ciiT) ■^ CO CO CO

O O yD CI C(



eg 1^


0-2 '^ "Phco " "

^ 0-7^T3 « 


m lo^d t^ lO lO'Xi t^ t^t^ t^ t^to ^

o >p

01 ►H

r^ « O ^

M CTi CO lOCO f^

« 6 COti> 131 "

CO CO CO CO c< c<

w O

CD lO CO ^ lO lo in t^

o CO a> Ti-co lo CO ■* Tj- CO o

O ■* 1^ CO r^cp LOCO ^ CO t^ ^ ■^co r^ 7

M^r^OCOi-CTiLO rhcb CTi f^ f^ f^ CT) ->


o O

'-30331'iU -^aj

QOhOKffi K

O «  . C

> y


CO " O O lO lO CT) en lO C"!

6 6 ci 6 io

N " " c< "

« „' en

O « ^ « 


o E ~'- X^


o f^

u ^— '


s a


1/3 o

c ^ c _

CO 7*" r^ O

1X> ^ C5 CT)

CO ^ O K)

6 6 « 6

in ■* irs jj

o in o


(i ti t£i

O) 6 Cl CD ^ uO Tf ■*


o o o o

O -1 O CD

(i> f^cb f^

lO lO lO lO


3 O




bog g

C C u O fc^ ^^ aj

i3t)cubDl'-3 a



be 3 u


2 K


o be

E ^


2 -S



old-fashioned terminology. Gobley made many quantitative observations on the various substances, and his figures are given in Table 17, which sums up all the analyses that have been made of the yolk in the eggs of birds. The original discovery of vitellin was made by Dumas & Cahours, but Gobley was the first to make an extended study of it. His elementary analysis is given in Table 10 a. He knew that it contained both sulphur and phosphorus. Gobley was able to isolate oleic and margaric (palmitic) acids from the fat fraction of the yolk, but, unlike Planche twenty-five years before, he got no stearic, and Kodweiss, one year later, reported its presence under the impression that it had not been found before. Gobley, however, was easily able to repeat Lecanu's discovery of the presence of cholesterol, and made a remarkable examination of the lipoids. "These viscous materials", he said, "appear to have been considered by John as not being of a fatty nature at all. They form the most interesting part of the yolk; they contain all the phosphorus which exists there in considerable quantity." He analysed the glycerophosphoric acid which he obtained from the lipoid, which he named "lecithin", and speculated as to the significance which it might have for the growth of the embryo. He also recognised that fatty acids and nitrogen were present in the viscous matter.

Ten years later Valenciennes & Fremy made a further examination of the yolks of a large variety of eggs with special reference to \itellin. It was they who discovered substances very similar to vitellin in the eggs of reptiles and fishes; these they named the ichthulins. As regards the eggs of birds, they contented themselves with confirming the results of the previous in\'estigators, but they regarded vitellin as having practically the same constitution as fibrin, on the grounds of elementary composition only. At the same time, they held it to be a different compound because it would not, like blood fibrin, decompose hydrogen peroxide.

If Table 1 7 is examined, it will be seen that the yolk is much drier than the white in all birds' eggs examined, having only about 50 per cent, of water as against the 85 per cent, of the latter. On the other hand, the percentage of fatty substances and lipoids is much higher, being just about double the amount of protein, whether related to wet weight or to dry. It is noticeable from the analyses of Tarchanov that the yolks of eggs from nidicolous birds having a short incubation time are about 10 per cent, richer in water than yolks from the eggs


of nidifugous birds. This must imply that the greater requirement for nutrient material in the latter case has, as it were, packed the fat tighter into the yolk. Exactly the same relationship is brought out from the figures of Spohn & Riddle, who compared the pigeon which hatches out as a squab with the hen which hatches out as a fully-feathered chick. Spohn & Riddle's analyses are the only complete ones we have for a nidicolous egg, and bear clearly the same relationship, for there is less protein and less fat, relatively, in the pigeon's egg than in the hen's. The ash content and the amount of non-nitrogenous extractive substances seem, however, to be slightly higher in the latter case. Langworthy's figures were all obtained from the eggs of nidifugous birds, and they show a great similarity among themselves. More delicate consideration, of course, reveals differences according to breed in the hen's egg, e.g. the figures of Pennington and his collaborators, but these are of a comparatively minor order.

The most interesting analyses are those of Spohn & Riddle. They compared the egg of the jungle-fowl, which is supposed to have been the evolutionary ancestor of the domestic hen, with averaged figures for hen's eggs of various breeds, and, as is evident, there was a very close agreement. They also analysed the white yolk as distinct from the yellow yolk of the hen's tgg. When the yolk begins to be formed in the ovary of the hen, it is white and not yellow, and not until the critical point in its maturation is reached, when its growth-rate completely changes, does it begin to store lipochrome pigment. This change in growth-rate, which has been observed by other workers as well as Riddle (e.g. Walton), will be dealt with in more detail in the appendix on maturation. Von Hemsbach, in a paper on the milky or white yolk of the birds, in 1851, suggested that the corpus luteum of mammals corresponded to the yellow yolk of birds, and that the mammalian ovum having been shed out of the ovary into the Fallopian tube and uterus, the fats and lipochrome pigment were laid down in the Graafian follicle instead of around the white

'ovum". Von Hemsbach also supported the view already mentioned that the shells of avian and amphibian eggs corresponded to the decidua of mammals. He laid stress on the work of Zwicky and Gobel, who had investigated the pigments of yolk and corpus luteum, and had thought them to be identical. This subject will be referred to again under the head of pigments.




In the fresh egg, as laid, the white yolk occupies a central position, and is surrounded by concentric layers of yellow yolk. But as a kind of cylindrical prolongation of the white yolk reaches to the surface of the vitellus underneath the blastodisc or germinal spot, the white yolk must be considered the first food of the embryo, and, until its composition was determined, it was not possible to say what sort of nutrient environment the embryo possessed in the very early days of development, although the composition of the yellow yolk would give this for the later period. The histological differences between white and yellow yolk had been known for a long time (see Purkinje ; His ; Leuckart ; Klebs ; Dursy ; Strieker ; and Virchow) but Riddle and Spohn & Riddle were the first to approach the subject chemically. Their figures showed that the white yolk much the more nearly approximated to the contents of invertebrate eggs with holoblastic cleavage, and living undifferentiated tissue generally. Instead of 45 per cent, of water, the white yolk had 86 per cent., instead of 15 per cent, of protein, it had only 4, and instead of 25 per cent, of fat it had only 2. Thus in its water-content, it was much more like {a) eggwhite and {b) the young embryo itself than like ordinary yolk, while instead of having twice as much fat as protein it had twice as much protein as fat. These data are extremely interesting in view of the facts that are known about the sources of energy made use of by the embryo during its development. Although by far the greatest proportion by weight of substance combusted during embryonic life is fat, yet, in the early stages, the embryo undoubtedly gets its energy preponderantly from protein and carbohydrate (see the whole of Section 7). The percentage of non-nitrogenous extractives did not differ much between white and yellow yolk in the experiments of Spohn & Riddle, but it would be very interesting to know the relative amounts of carbohydrate, and analyses to discover this should certainly be done. Again, the yellow yolk contained eight times less ash than the white yolk, a finding which acquires considerable significance from the fact that, if the ratio inorganic substance/organic substance in the embryonic body is plotted, it is seen to descend steadily from the beginning of development (see Fig. 249). Moreover, as Mendeleef has shown, early embryonic cells contain twice as much electrolyte as those of later stages (see Section 5"8). The amount of phosphatide in the yellow yolk, furthermore, was ten times that in the white, a significant difference; for, as Plimmer & Scott have shown,


one of the main functions of the phosphatide is in furnishing phosphorus for the embryonic bones during the period of ossification, a requirement which is not present in the earher stages of the development. The histochemical work of Marza, who compared the white and yellow yolk following the method of Romieu, is in agreement with this, for he found the elements of the yellow yolk to be richer than those of the white. (See Plate X.)

1-8. The Avian Yolk-proteins

As regards the protein, vitellin (Tables 10 a and 11), several interesting points are to be observed. The best elementary analyses of ovovitellin are probably those of Osborne & Campbell. After its discovery by Dumas & Cahours, Gobley, and Valenciennes & Fremy, it was studied by Hoppe-Seyler, and now for the first time with special reference to its position in the classification of the proteins. Virchow had some time before then suggested that the yolk-platelets, familiar to histologists, contained lecithin, and there had been some doubt as to their nature. Valenciennes & Fremy had opposed the view that they were crystals, basing their view on Sennarmont's work, but Radlkofer and Hoppe-Seyler returned to the crystal theory. Hoppe-Seyler believed that vitellin contained no phosphorus, but that what appeared in the analyses was due to contamination with lecithin. This view was supported also by his assistant, Diakonov, who contributed to the Med.Chem. Untersuchungen one of the earliest investigations of phosphatide. But at the same time Miescher obtained from the yolk of the hen's egg a substance containing a great deal of phosphorus, and possessing certain of the properties of a protein. This he believed to be nuclein. "It is interesting ", he said, "in relation to the origin of nuclear substance, that the nutrient yolk contains ready-formed nuclein in significant quantity." At this time, then, the proteins of the yolk were believed to be ovoglobulin (for so Hoppe-Seyler called the vitellin of the earlier workers) and Miescher's nuclein. Miescher himself identified his nuclein as a constituent of the white yolk of the histologists, but he noted that the hen's egg seemed to have no xanthine in it.

Lehmann, Schwarzenbach and others, however, did not agree with this classification, and regarded vitellin as a mixture of albumen and casein. They did so not on the grounds of its containing phosphorus, but because they found that rennin would completely coagulate it from its pure solution. But this attitude did not prevail.


and the word " nucleovitellin " became general, until Kossel in 1886 found that, if vitellin was really a nuclein, it differed from all other such substances by giving no trace of xanthine after acid hydrolysis. On the other hand, true nuclein, he found, was present by the tenth day of development. Hall and Burian & Schur, Bessau and von Fellenberg confirmed this absence of purines from the fresh egg. In more recent times, Sendju and Mendel & Leavenworth have found exceedingly small amounts of true nucleoprotein (2 and i*6 mgm. per cent, respectively wet weight) in the hen's egg (by purine bases), and Plimmer & Scott, and Heubner & Reeb have done the same (by phosphorus analysis) . Shortly after Kossel's work, Milroy found that vitellin gave a biuret test though no Millon, and materially differed in nitrogen and phosphorus content from any of the nucleoproteins, while, at the same time, Miescher admitted that he could not isolate any purine bases from his "nuclein" in the hen's egg. Levene & Alsberg next investigated the manner of breakdown of vitellin, finding the substance they named " paranuclein " after digestion with pepsin, and "avivitellic acid" after hydrolysing with ammonia. The elementary composition of these substances is given in Table 10 a, from which it could be seen that the increasing phosphorus content implied the presence of phosphorus as an important constituent of the original molecule. Six years later Levene & Alsberg ascertained the amino-acid distribution (see Table 11). They pointed out the significance of the high proline figure, in view of the task of haemoglobin synthesis which the young embryo has before it. Abderhalden & Hunter and Hugounenq undertook a like investigation in the same year, from which a striking similarity between the amino-acid distribution in vitellin and casein came to light, especially as regards the high proportion of leucine and glutamic acid. They drew attention to the similarity in physiological requirements as between the "erste Nahrung" of chick and mammal. The historical associations of this discovery have already been referred to (see p. 53). It was at this time that Neuberg, and Blumenthal & Mayer reported the existence of glucosamine in the vitellin molecule, two observations which stood together in isolation, until in 1929 Levene & Mori isolated from egg-yolk the same trisaccharide which they found to be present in ovoalbumen and ovomucoid and which has been referred to above.

It was not until the paper of Bayliss & Plimmer in 1906 that the



/ kj^EHHHH^^^HHI

' '^ "


Stain, haemalum-eosin: magnification, gxA: prepared and microphotographed by Dr V. Marza. The stratification of the yolk into white and yellow is beginning.


nature of vitellin really became clear. They subjected casein and vitellin to the action of trypsin, and studied the time taken under varying conditions for the phosphorus to be split off into soluble form. Ovovitellin, they found, was much more slowly digested than casein, for after 36 days only half of its phosphorus had been made soluble, whereas after 2 or 3 days a large percentage of the casein phosphorus had gone into solution in inorganic form, and most of the rest was present in water-soluble organic combination, i per cent, soda, however, would bring aU the phosphorus of casein into solution in 24 hours. BayUss & Plimmer concluded that ovovitellin and casein were both phosphoproteins, as distinguished from nucleoprotein, where the phosphorus would be present in the prosthetic group and not in the protein itself Plimmer & Scott later found that ovovitellin behaved in the same way to soda. This reaction served to distinguish between phosphoproteins and nucleoproteins, for all the latter, it was found, were stable to alkali and easily split by acids. From the phosphorus distribution in the unincubated hen's egg, Plimmer & Scott concluded that vitelHn accounted for at least 30 per cent, of the phosphorus, and this led them on to their investigation of the changes which take place in the different phosphorus fractions during the development of the embryo.

The distribution of phosphorus-containing compounds in egg-yolk, as Plimmer & Scott found, makes a very different picture from that of any other tissue. Their summary is shown in the accompanying table (18). It would be extremely interesting to investigate the phosphorus distribution in the white yolk, which at present is altogether uncharted.

Table 18. Phosphorus in per cent, of the total phosphorus.









Lecithin P ...





Total water-soluble P





Water-soluble inorganic P





Nucleoprotein P





Phosphoprotein P ...





Total protein P





The third of these fractions i

includes the phosph

orus of all unstable water-soluble corn


Steudel, Ellinghaus & Gottschalk have recently found that vitellin behaves towards pepsin exactly like casein. The rate of increase of titratable COOH groups during the digestion far exceeds that of NH2 groups, reaching a maximum about the fourth hour. The




[PT. Ill

Table 19.

Per cent, of dry weight

Total N



N/P rati

mer's figures.

Ovovitellin ... ovolivetin ... Ovoalbumen ... Casein ...

... 15-29 15-12

••• 15-51 ... 15-30

i-o o-i o-i

15-3/1 151-0/1


Levene & Alsberg's figures.

Avivitellic acid ... 13-13

Swigel & Posternak's figures.


Swigel & Posternak's figures.

Hydrolysis of ovotyrine b^ (% ) Pyruvic H3PO4 acid NH3 Arginine Histidine Lysine

12-00 1-60 4-90 0-62 0-70 0-75


Ovotyrine a^ ...

... 10-87




Ovotyrine ^^ ...

... 11-33




Ovotyrine /Sj ...

... 10-92




Ovotyrine 71 ...





/-serine 7-90

Table 20. Nitrogen distribution.

Plimmer's figures (1908).

Per cent of dry weight

Ovoalbumen Casein

Ovovitellin . Ovolivetin .



15-51 15-30 15-29 15-12




1-52 0-84 0-75


N 0-29 0-22 0-25 0-32



3-30 330 3-84 3-29

Monoamino N 10-58 10-36 10-26 10-76

linkages must break, therefore, between a carboxyl group and proline, tr-yptophane, histidine or arginine.

Weyl ; v. Moraczevski ; and Gross were the first to describe the properties of the egg-yolk proteins, but the standard account is that of Plimmer, who in 1908 identified two yolk-proteins, ovovitellin, and ovolivetin. Ovovitellin, according to his analyses, contained 1*0 per cent, of phosphorus, but ovolivetin only o-i per cent. He was usually able to isolate far more of the former than the latter, but in some experiments the yield seemed to be nearly equal. Livetin was

SECT. l]



soluble in water as well as 10 per cent, salt solution, but it cannot be ovoalbumen or any of the egg-white proteins, for it is not coagulated by ether. Plimmer suggested that possibly livetin was vitellin with the majority of the phosphorus-containing parts split off from it. Tables 19, 20 contain Plimmer's figures for these two proteins.

Table 21.


May & Rose

Folin & Looney

(1922), (%)

(1922), (%)





















Free tryptophane (% of whole egg)

von Fiirth

& Lieben

Ide (1921)


Whole egg-contents






0-437 "otal tryptophane.

(% of proteins)

(% of proteins)

Whole egg-contents








Whole egg-contents


Ovovitellin has been the subject of recent investigations by Swigel & Posternak. They found that it broke up into three polypeptides which they call ovotyrine a^, ^^ and y^. The properties of these are listed in Table 19. It was found that ovotyrine ^ contained all the iron in the original compound, and that it could be split up into ovotyrine /Sj and ovotyrine ^2 the second of which again contained all the iron. All these derivatives were laevorotatory, and showed considerable resemblance to the lactotyrines which the same workers had previously isolated from casein. They identified their ovotyrine ^ with the avivitellic acid of Levene & Alsberg, and they stated that an enzyme was present in the fresh yolk which would, on standing at 37° C. for 10 days, double the yield of preformed ovotyrine jS.


— - u

O 3

5P =0 § a







•ft IN



orjouii % I

XxojpAq % I

DUBajs % I

oijnuiBd % I

oiajo % O


oijAjnq % I

3% I

N% I

•3 -ds I

xspui aApoBija^j | •ou jauqaH ^^

•ou ]A}30V I


anjBA jajsg r~


an^BA ppv ?

•ou uopBogiuodBS


in CO

O) o


aim oo

•ou auipoj

coco o -^ p eo^ oj w Tho^eoco Tt< e« fj " a f^ ^ii> N tJ< 6 f^ o» 'X> (^ t^co lo cotn i^co r^ c< <J) lO

in o coco (ij 6 CO 6 CO t^UD i^

c fl S^ ci o o

< -C(V C^tH in

•3 "^ u . "^ <« ?4

^ D o5i ° n£,-£.

iH Pffiffi

2 2^ &• ^


g jj o s;a cn

c c fi 2 c 2 3



^ « « ^ a

^ o o

'^ - - •< « 





s i


e 3 S3 w


v5 «^

s >- p fl

pHC/2y3 3H













O o


O f~~ CO mCO CT) ^ ■*C0 iD

CT:cb w c< 6 '^ •4'6d ^ CO (£1 •*co r^ r^to i~~ i^ CO (j2


^ bo "^

.ii.a a i^z

c« oj y u c &. c

^ ci, u X! :r! u=! *■

a c c c c c V

o o o o o o


g C3 g

c £ « • o V o a

"* c e

o o




Hydrolysis of ovotyrine ^ revealed the presence of large amounts of /-serine, an amino-acid which had not previously been found in ovovitellin (see e.g. Plimmer & Rosedale's analyses). Some pyruvic acid and ammonia being given off as well, Swigel & Posternak calculated that, supposing these arose from breakdown of serine, there would have been sufficient serine present initially to combine with for all the phosphorus. They therefore suggested that the main phosphorus-containing unit of ovovitellin was serine-phosphoric acid.

Cohn, Hendry & Prentiss consider the minimal molecular weight of vitellin to be 192,000, i.e. much higher than ovoalbumen.

Kay & Marshall have also studied the yolk-proteins. They have prepared purer samples of vitellin and livetin than those of any previous worker, and have been able to free the former almost entirely from contamination with ovolecithin. Their vitellin is a true phosphoprotein containing i -3 per cent, of phosphorus and hydrolysed by I per cent, soda, though not by the phosphatase of the kidney. Their livetin is a pseudo-globulin, containing only the slightest traces of phosphorus (less than 0-05 per cent.). The yolk of the fresh egg contains no albumen. Vitellin, hydrolysed with dilute ammonia, gives a vitellinic acid containing about 10 per cent, of phosphorus. Kay has estimated the cystine, tryptophane and tyrosine in vitellin and livetin (Table 11); in the latter protein they are distinctly high in amount, a fact of some importance in embryonic nutrition. The relative amounts of vitellin and livetin in yolk would appear to be of the order of 3-6 to i for the hen and 3-8 to i for the duck, calculating from their nitrogen content. Kay regards livetin as identical with Gross' protein. The yolk of a fresh egg would contain from 600 to 900 mgm.

1*9. The Fat and Carbohydrate of Avian Yolk

The fatty acids of the yolk have been much investigated since the time of Gobley and Kodweiss, but little has been added to our knowledge of them. Paladino found olein, palmitin and stearin to be present. Analytical details are in Table 22.

A large part of the study of phosphatides, under the generic name of lecithin, has been made on that obtained from the yolk of the egg; thus the work of Diaconov, who showed it contained no neurine, Strecker, who discovered the presence of choline, Bergell ; Cousin ; Laves & Grohmann; Laves; Wintgen & Keller; Erlandsen; Stern


& Thierfelder; Frankel & Bolaffio (whose egg-yolk neottin was only a mixture of sphingomyelin and cerebrosides), McLean; Serono & Palozzi; Eppler; Riedel; Wilson, and Trier, who prepared aminoethylalcohol from it, all comes under this heading. In McLean's book will be found a review of it. Certain aspects of it, however, are important here ; for instance, the question of the presence of very unsaturated acids in ovolecithin. McLean in 1909 found stearic and oleic acids in it, but Cousin was able to isolate linolenic and palmitic as well, and Riedel; Hatakeyama; and Levene & Rolf obtained linolic and arachidonic acids. In another paper Levene & Rolf showed that the lecithin, carefully freed from kephalin, contained only palmitic, stearic, and oleic acids : saturated and unsaturated molecules being present in equal proportions. Again, Stephenson in 1 9 1 2 found an acid in the phosphatide fraction from egg-yolk, which had 20 carbon atoms and 6 or 8 unsaturated linkages. Although the proportion of unsaturated acids in egg-yolk is generally agreed to be small, yet it may be of importance for the young embryo if it passes through a period in the early developmental stages before it has the power of desaturating the ordinary fatty acids. Evidence which suggests this will be presented later (Section 1 1 • i ) .

The nitrogenous radicle in ovolecithin is largely choline, but difficulty was at first experienced in obtaining a theoretical yield on hydrolysis; thus Moruzzi got only 77 per cent, in 1908 and McLean only 65 per cent, in 1909. This was accounted for, however, when it was found that amino-ethyl alcohol was also present. The two bases together make up all the nitrogen in the molecule. Erlandsen was the first to question the view that lecithin alone accounted for the phosphatide fraction, but he was not himself able to isolate anything else. Later workers (Levene & West and Stern & Thierfelder), however, found that kephalin is also present in yolk, and it would probably be in the kephalin molecule that the unsaturated fatty acids would be present. Analyses of kephalin from the yolks of fowls are given in Table 10 ^. McLean in 1909 isolated from egg-yolk a third phosphatide which resembled cuorin, but it is very doubtful whether this was a true chemical individual. Sphingomyelin has also been found in egg-yolk by Levene (191 6), and lignoceric as well as hydroxystearic acid was present in it.

All these substances exist in the yolk in close association with the proteins. Hoppe-Seyler it was who first observed that, after


prolonged extraction of the yolk with ether, a considerable proportion of the phosphatides still remained behind, and could be extracted with alcohol. It was thought for a long time that the phosphatides and the vitellin were in chemical combination which was broken by the alcohol, but since the paper of Fischer & Hooker in 1 9 1 6 the general opinion has been that this combination is only physical. Stern & Thierfelder isolated traces of the cerebrosides, phrenosin and kerasin, from egg-yolk in 1907.

The neutral fats and the lipoids of the yolk are variously affected by the nature of the fats in the food of the fowl. Henriques & Hansen, who were the first to investigate this subject, found that, if food containing very unsaturated acids was fed to the laying hens, the neutral fats in the eggs were affected, but not the fatty acid components of the lecithin. Their figures are shown in Table 22. When the food consisted of barley, pea or rice, the iodine number of the neutral fats in the egg varied round about 77, but hemp or linseed sent it up to about no, although no matter what the food was the iodine number of the fatty acids in the phosphatide fraction remained constant at 75 or so. Henriques & Hansen also found that the iodine number of the fluid fatty acids of the neutral fat was normally 107-5, and that the fluid and solid fatty acids of the phosphatide fraction were 151-3 and 98-9 respectively. The former accounted for 64-3 per cent, of the lecithin fatty acids. The experiments of Henriques & Hansen have been repeated and confirmed by Belin and by Terroine & Belin. The last-named workers, together with McCollum, Halpin & Drescher, some years later reported that the lecithin fatty acids would vary, as well as the neutral fatty acids, with the diet of the hen. Their figures, which are given in Table 22, certainly show a variation in the iodine numbers of both fractions. All these workers recognised the presence of unsaturated acids in the yolk fat, and Henriques and Hansen's figures came between the theoretical values for oleic and linolic acids.

Work was continued along these lines by McGlure & Carr. Using pigeons, they found that the fat content of the eggs could only be altered slightly by feeding rations high and low in fat.

% fat in the eggs Cocoanut fat ... ... ... 4-0

Beef tallow ... ... ... ... 6-75

Average of all fat diets ... ... 4-96

Average of all non-fat diets ... 4-81

3 be

■^ ^


be ffi


















C5 •:


(U bO S


^ OUIUIB l^jox

^ uiajojd ajqEinSBOo



|«j^ upjojd 3iqBinSBOD-uo]y[

N F50X


^ ouiuiB aaaj

^ UT3}Ojd

^ uiajojd ajq^inSBOD-uofj

fvj OUIUIE 331^

fy[ asouinqiY



\[ uiajojd 3{qB|nSB03

I^J upjojd-uou 3{qB{n§BOD-uo^

|vj uiajojd 3iq^|n§BOD-uo^



bo !^

•j3 "O










tr, u:,





!r> o


o o o

O ^ IN ir^ lO 1^^

OO ir>OoO,oOO COtXJ ,O^0D IT) O <0 LO CO

' « «" <N «'" -^ kT ;^ «" cT




V •

^ s

bo w

• O

e &

be V

S I:

"O S c « 

o >^

^ ^-^ S^.^


^ g § ^



[PT. Ill

Again, during the fat feeding, tiie saponification value of the eggfat was 176 (166-190) and during the rest of the time 173. The iodine value was in the former case 70-5 and in the latter 70-8.

Table 24. Lipoid in egg-yolk.




/o /o Wet weight Dry weight






















Glikin' s figures.

Total fatty

P,05 in Lecithin




% of in %



% dry

fatty of fatty




acids acids


Pigeon (yolk)



3- 16 35-73




3-88 38-42


Turtledove (yolk)



4-10 46-65


Starling (whole egg) ...


5-67 64-44


Hen (yolk)


— —


Thrush Cat ... Rabbit Guinea-pig

Lecithin in % dry weight at birth or hatching 8-18 5-06 4-91 3-79

Some suggestive investigations on the biological significance of ovolecithin were made b-y Glikin in 1 908, whose figures are shown in Table 24. Choosing the pigeon as a typically nidicolous bird, and the hen as a typically nidifugous one, he was able to show, using a variety of extraction methods, that the yolk of the former was considerably richer in lecithin than the latter, the former containing about 29 per cent, dry weight, and the latter about 17. The further but rather fragmentary observations which he made on the starling and the turtledove confirmed this relationship. It is interesting that Tso informs us that certain small Chinese breeds of hen produce very small eggs (scarcely 40 gm.) and that these contain a much higher percentage of lipoids than ordinary eggs though an equivalent percentage of protein. He concluded that lecithin, one of the most essential yolk-constituents, was specially concentrated in nidicolous yolks and


was associated with the property of early hatching or birth. Thus he compared the thrush (nidicolous) with the guinea-pig, which is born almost ready to eat green food and hardly passes through a lactation stage; in the body of the former 8 gm. per cent, lecithin dry weight was found, in the latter only 4. The new-born cat and rabbit occupied intermediate positions. It is interesting to note that Glikin's figures bear out those of Tarchanov on the question of water-content of yolks from the two types of birds.

Tornani affirmed in 1909 that a difference in lecithin-content was observable between fertilised and unfertilised eggs. But as he gave no figures in support of his contention, it has not been treated with much respect by subsequent workers.

The carbohydrate of the yolk has been the subject of only a very few researches compared with that of the white. The figures which have been obtained are shown in Table 16, and it will be seen that in no case has the total carbohydrate been estimated, and only in one case the glycogen. After Claude Bernard's isolation of glycogen from the yolk, a persistent belief grew up that considerable amounts of this substance were present there ; this was apparently based on the description by Dareste in 1879 of "amyloid bodies" in the yolk which gave microchemically a strong blue colour with iodine. Dastre immediately pointed out that the occurrence of starch there was highly improbable, and that if any glycogen was there it should give a wine-red colour; he himself, however, could find neither. But he did not succeed in suppressing the rumour, for Virchow, and later Schenk, supported Dareste, though nothing has been heard of this yolk-constituent since 1897, and Sakuragi's analysis revealed the presence of only 2-2 mgm. per cent, of glycogen. Bierry, Hazard & Ranc asserted in 191 2 that they could obtain a great increase of carbohydrate after hydrolysing the yolk with hydrofluoric acid under pressure, but this would not imply, as they seemed to think, that glycogen was present, for all kinds of other compounds such as proteins (Gross' protein for instance) might yield glucosamine under such treatment. They identified glucosamine in the hydrolysate. On the other hand, Diamare, who hydrolysed with acetic acid, could only obtain faint traces of combined glucose in the yolk. He dialysed both white and yolk, and in both cases was able to estimate the free sugar, but in the case of the yolk very little combined glucose seemed to be present. Further studies on this


subject should be undertaken, for the methods of Diamare and Bierry ahke were of questionable reliability. Diamare, however, went rather further into the matter than other investigators, and, thinking that the yolk glucose might only be present there owing to an inflow from the white, examined the ovarian eggs, in which he found glucose in much the same proportion as in the yolks of laid eggs. He does not state whether the ovarian eggs were yellow or white, and, as he frequently gives his results in the form of grams of glucose without mentioning the weight of the fresh material, it is impossible to calculate the percentage (see also Tillmans & Philippi) .

We have already seen that cholesterol was identified in the yolk of the hen's egg by very early workers such as Gobley. In 1908 Menozzi and in 191 5 Berg & Angerhausen sho\ved that egg cholesterol was identical with that from milk and bile. It is certainly present in the unincubated yolk both free and in esterified form with fatty acids. Serono and Palozzi investigated a substance from egg-yolk in 191 1 which they called "lutein" but which turned out to be nothing but a mixture of cholesterol esters. Other investigators have estimated the amount of free and combined cholesterol in the unincubated egg, and their figures are given in Table 25.

Table 25. Cholesterol-content of hen^s egg.


per whole egg

Investigator and date

Parke (1866)

Mendel & Leavenworth (1908) Mueller (1915) ... Ellis & Gardner (1909)... Thannhauser & Schaber (1923)

Cappenberg (1909)

Dam (1929)

Schonheimer (1929)

Cholesterol Cholesterol (free) esters

378 215-9 24-2 489 173 54-2

296 — 337 —

Total 248-3

240-1 600-6 227 180

Free in % of total





Many Other substances have been found to be present in the egg at the beginning of development, e.g. choline, alcohol, creatine, creatinine, inositogen, lactic acid, plasmalogen (Stepp, Feulgen & Voit, 1927), etc. These will be mentioned as occasion arises during the succeeding sections of the book. Allantoin is not present (Ackroyd).

The yolk of the hen's egg also contains vitamines, pigments, and a variety of enzymes, but these will be dealt with under their respective sections. As Langworthy has shown, it may also contain very

'§,0° ? CO V C -!i; Ci cfl an

S: o' " ?^M ^ ?

o o>^


en M O

C — O"

5 2 e5 c9 fc,

08 O'c


E 0)

s a " u

-§ S f^a• 0^^ ^ mi

ja :'

"ego c

>^4 ::

d uispnjsj

d aiqnios


d ajqnios J aiqnios

J upjojd -oqdsoijj

d appEqdsoqj

d unPMA

d ajqnjos


jaiqnios -J31BM 3UIe3jo So I

d aiqnios "I

-J31BM pjOX, g f

d siqnps -loqooiB iBjox ^

(uonoB.ijxa ^ jaqja iajje) d r ajqnps-ioqoofY

d aiqnios-jaqja |



d aiqnps n


d aiqnjos r -ja^BM 3jub3jo

d aiqnios | -jajEM iBiox

d siqnios


(uotjoBjaxs 7^

jaqaa iajjE) d j^

ajqnjos-ioqoDiv ip

d aiqnps-jsqjg ],

d lElox

t^ m >^ iri v^o O O

O (U

^ >. -•

^'o "u

^>> as

j: a


Oj: ■Co o-g

^^ ^^

(U u T3 « 

C O 3 ^

5 °

•^ D.

a 1

m (U

o c

J3-Q.2 V C

3 ao

■+ir) 000





various substances derived from the diet of the hen, and these, if they are odorous or possess taste may very easily betray their presence (e.g. the Swedish "Schareneier" described by Hansson). Table 23 gives the figures which are available for the nitrogen and Table 26 for the phosphorus distribution. These summaries bear out on a detailed basis what has already been said.

I -10. The Ash of the Avian Egg

The ash of the yolk and the white of the hen's egg has been investigated by several workers, and a study of it reveals certain interesting features. If Table 27 be examined, it will be seen that, in the yolk as well as the white, potassium has almost invariably been found to be present in greater amount than sodium. This is one of the characteristics of the egg-cell, as will be seen later when the eggs of other animals are considered. The yolk is also marked by the very high percentage of phosphorus, most of which is, in accordance with other evidence, in organic combination. The calcium is also mainly in the yolk, as is the iron, but not the magnesium. If now the amounts of metallic and acidic ion be calculated out in millimols and milliequivalents per cent, wet weight, it is found that in both yolk and white there is an uneven balance, but while in the former case there is much more anion than cation (anion/cation ratio above unity), in the latter case the exact reverse holds, and the anion/cation ratio is somewhat below unity, about 0-55. In the white, therefore, some of the potassium and sodium must be combined with the proteins, as ovoalbumenates, etc.* However, the excess of cation over anion in the white is not so considerable as the excess of anion over cation in the yolk, and, bearing in mind also the much higher percentage of solid in the yolk than in the white, it would be expected that the anion/cation ratio of the whole egg would be greater than unity, and would approach that of the yolk. The facts show that this is, indeed, the case, for the average anion/cation ratio calculated from the results of all observers for the whole egg is 2-3, as against 2-8 for the yolk alone and 0-54 for the white alone. This was first noted by Garpiaux. Forbes, Beegle & Mensching expressed it simply thus :

Cubic centimetres normal solution per 1 00 gm. dry weight egg Total acid ... ... ... 120-28

Total base ... ... ... 39*42

Excess acid over base ... ... 8o'86

In both white and yolk, of course, the inorganic ash is basic.

Duck Hen ...


Table 27. Aili of the avian


% of total ash








PO, c












79-1 —






Bi-2 15-2

Mgm./loo gm. wet weight

K Na Mg Ca

Whole 2-59 2-72 0-75 3-28

)-33 048 0-79 3-5

Fe SO, PO,

.,.„. . , Total cation Total anion

Millicquivalents . '■ ^ , * ^ Anion/

K Na Mg Ca Fe SO, PO. CI mols equiv. mob equiv. ratio Investigator and date

3 — 2-59 2-72 1-50 6-56 — — 30-9 — 9-34 ,j.j7 ,0-3 jo-j 231 Polcck (1850)

y - 3-33 0-48 1-58 7-u - - 35-7 - 8-t ,2-39 ,,-9 jj.7 ,.-88 Rose (1850)

65 738 4-5 0-9. 1-66 7-0 — _ 34-95 7.38 9-79 /.,■„ ,9.03 ^.^j 3.00 Bialascewicz (1926)

3-32 50-3 S'll" 173 204

93 — 3-8 604 ]o8 443 8-8 0-34

004 636 302 443

4-6 — o-o8 19-08

S-j/ 9-42 s2-;a

Plimmer & Lowtidnt (1927) Vaughan & Bills (t878) Delezeime & Foumeau (1918)

Carpiaux {1908) Buckner & Martin (

Normandy Bourbonnais

)125 17-5 273 fo

)I30 15-4 23-8 Trace

3-590 23-55 ao-7 0-96

0-720 a7-tiG i'2-i 2-7

0-634 — — 24-B 25

o-Goo 22-1 20-63 1-41 I
















0-52 2-9 1




4-87 While




0-24 o-.e

■ 2



o-Bi 052

0-65 0-40

3-56 5-3 5-11 3-8









7-17 77-0^ 0-C25 Champion & Pellet (1876)

(not very reliable) 7-77 17-84 0-G29


0-9 4-4 9-39 s-s8

S-72 0-58 Poleck (1850)

3-51 4-12 035 Rose(iB5o)

- - - Voit(,877) 4-82 ssf c-55 lljin (1917)

— — — Prout (1822)

— — — — 4-25' —

523 4-l'8 3-79 1-24 8-72 345 '0-9 '-"8 2-2 13-62 —

129-a 52-5

85-5 4-25 Go- 16 3-6 84-6 —

64-7 1-8 48-4 3-7

.■58 6-6 !-48 9-0 2-4 9-6

,2-48 13-81 12-69

Average ... 0-54


- - 139-8

Straub & Hoogcrduyn (1929)

Poleck (1850)

Rose (1850)

Voit (1877)

Carpiaux (1903)

lljin (1917)

Prout (1822)

Straub & Hoogerduyn (1929)

Kreis & Studinger.


This is probably the most interesting consideration that emerges from Table 27, but it may also be noted that the ash-content of the white is just about half that of the yolk, a relation which would practically be reduced to equality if the phosphorus in the yolk was not taken into account.

The presence of certain chemical elements of lesser biological importance has been announced from time to time in a group of papers which have some interest, although it is difficult to see, as yet, what their importance is for the development of the embryo. Fluorine has been estimated by Tammann and by Nickles, copper by Dhere, boron by Bertrand & Agulhon, manganese by Bertrand & Medigreceanu, iodine by Bonnanni and by von Fellenberg, lead by Bishop.

These elements appear to be normal constituents of the egg. The iron-content can be artificially increased by feeding iron-rich rations to the hen, and iodine can also be introduced into the egg in this way, as has been done by Bonnanni, Kreis and others, but Hofmann found that though iron and iodine would enter the egg thus, it was impossible to get copper to do so. In just the same way Ricci found it difficult if not impossible to get As or Hg into the hen's egg by feeding subtoxic doses to the hen. The normal copper-content of the hen's egg cannot be varied like its iron-content. The importance of iron in the formation of haemoglobin is obvious, and the little that is known about this process will be discussed in the section on pigments in the embryo. Wassermann made a histochemical examination of the egg-yolk and vitelline membrane for iron, and found a relationship between the embryonic blood-islands and the iron of the yolk.

Some of the other data in Table 28 call for comment, Tammann's 1-13 mgm. per cent, fluorine in the fresh yolk works out at a quantity of 0-2 mgm. per egg, and, as Zaleski found 0-23 per cent, fluorine in the bones of the chick at hatching, o-o8 mgm. fluorine would be required in the egg at the beginning, or less than half of what is actually there. Zaleski's figure, however, is old, and may be too low. It would appear, on the whole, as if the greater part of the fluorine, iodine, copper, zinc, lead, aluminium, silicon and manganese is localised in the yolk, and the greater part of the boron and arsenic in the white. In view of the importance which we now attribute to these less common elements as catalysts in living tissues, this distribution may be found to have considerable significance.




Pi ^



























^ ^

.« "

^._, en «J 3 <^ >-i ir!

^ c «e ^

O^ o



„ a


.23 w

w 5<^

3 -"

_>. « 


"rt CU




o . , bD



SU > m

-d „ -0-3

o o o o

c ^

F « 

S bD bo

C; V3 C3



O CO ^- o o •■

?? "

-. CI " O o r-» o o o o o o o 6 6 6 6 6 6 6

5 =a


^ s

^..— V

c^ S

.-, lo

1-1 v!l CI

--^ f. W

c>! nj'^

Elvehjem Halpin Wolff (19 Lindow,


Fleurent i McHargu Bodanski




.-S-S 1

6 6 oi 6

? I


h K

-03 OH


S 2








^ ^

,. ^








W n




i-4 « 








„, — ,. '- ^



cn r.






CTi ^ ...





o - >-■ »3


S-i '•

J, ci

«^ it: ^ rt





^^ S-^

Oh cq M

6 6

« cq cq>Mm CLiOiiq Q

^ ^ ^ Oi ^ o





o o




(UN-" HP^

^ o -^ = ^0^:2. o _C ^ '^ '-I — '

P ^ c^


6 Oh

•£ bo

?2 o


tj t?




u ^

6 6 - 1



6 6




6 u ^ ■ — ^ "In. S3 2? J3 -^ ^ I I ffi ffi K S 3 S ^ < ■^ B c3 c S —.0.2 ^ V 1;; O bCc ■w o ^ S c w bo JJ b tP. ^■M U 3 4->

-> u c4 c S-5 o ex ,„ c_> n 5" .U_0 t, « 5 .§■« 3 fci ■" CX bo c -3 3o6 THE UNFERTILISED EGG AS A [pt. iii The figures which have been obtained by those investigators who have examined the iron-content of eggs are seen in Table 29. All found a great deal more iron in the yolk than in the white, as might have been expected from the earlier micro-chemical researches of Tirmann and Kobert. This kind of work had been originated by Schmiechovski, and was continued later by Wassermann in the interesting paper already referred to. Schmiechovski found iron histochemically throughout the yolk, but considered that, in the white or milky yolk, it was confined to the megaspheres. Table 29* Iron in hen's eggs. Italic figures indicate dry weight data. FcaOs gm. % wet weight Without iron-rich diet With iron-rich diet , ' ^ r ^ ^ Egg-white Egg-white Egg- plus Egg- plus white Yolk yolk Shell white Yolk yolk Shell Investigator and date •0024 -0088 -0047 -0272 -0040 'oogs -0059 -0272 Loges & Pingel (1901) — — -0046 — — — -0040 — Kreis (1900) •0057 -026 •0165 — — — — ^ Lebbin (1900) ■03 -05 -03 _____ •001 12 -00995 '00425 — — — — — Hartung (Mar. 1902) •00087 -01106 -00451 — — — — — ,, (May 1902) — — — — -00208 -01621 -00729 — ,, (June 1902) Trace -0108 — — — — — — Bunge (1892) — -0121 -0018 — — '0175 -0032 — Hofmann (1901) — -;r -0057 — — — — — Boussingault (1850) — -063 -og§ — — — — — Leveque & von Tschermak (191 3) None 'OI43 — — None '0143 — — Elvehjem, Kemmerer, Hart, & Halpin (1929) Wassermann, using both the ammonium sulphide and the Berlin blue methods, decided that it was present in both kinds of yolk, but that it was not confined to those special elements in the white part. In fact, it was very much more abundant in the white than in the yellow part. This finding has never been corroborated by chemical analysis, but, if it is, it will have considerable importance, in view of the time at which haemoglobin is most vigorously manufactured by the embryo. For a further discussion of these questions see the section on pigments. i-ii. General Characteristics of Non- Avian Eggs With this we may conclude the discussion of what is known about the typically terrestrial egg, that of the bird. Now SECT. I] PHYSICO-CHEMICAL SYSTEM 307 aquatic species far outnumber the terrestrial ones; as Spenser put it: O ! What an endlesse Worke have I in hand To count the sea's abundant progeny, Whose fruitfull seede farre passeth those on land, And also those which wonne in th' azure sky: For much more eath to tell the starres on hy, Albe they endlesse seeme in estimation, Than to recount the sea's posterity. So fertile be the flouds in generation, So huge their numbers, and so numberlesse their nation. It might therefore be supposed that a much greater space would have to be devoted to their eggs than what has already been taken up, but this is not the case, for the bird's egg has been so convenient a material for research that the knowledge we have of it outweighs that of the eggs of all other animals put together. Indeed, the data about the eggs of other groups are very fragmentary, so that much caution has to be used in making comparisons, and general relationships are much more difficult to enunciate. Van Beneden's classical memoir may be recommended as an account of the morphology of the eggs which are to be mentioned, and it is hardly necessary to refer to Balfour's book on comparative embryology. If Table 30 is examined, and compared with Table 2, it will at once be seen that the percentage composition of eggs of other classes of animals differs markedly and in very definite ways from the egg of the hen. The case of reptiles may first be taken, as being less remote than others. The reptilian egg seems to be distinctly drier than that of the bird, by about 20 per cent., and much more variable in its fat/protein ratio. For, while in all birds' eggs that have been investigated, the amount of protein, whether related to dry or to wet weight, is about the same as that of fat, the reptilian egg shows big variations from this rule. In the eggs of the tortoise and lizard, for example, there is twice as much protein as fat, while in those of the grass-snake, studied by Galimard, there is three times as much fat as protein. This fact will be mentioned again later (see p. 313). Considerably more is known however about those of amphibia, which have also been found to contain a great deal more protein than fat. Thus, instead of the 40 per cent, protein and the 40 per cent, fat which make up the dry substance of the hen's egg, Faure N " 3-— -_-N ■5 ^M 'o >^ 2 S C ■ H £ 3 K Own 53 fc 3 rx t c -a t

--- o o i-S «  ~«3 O O CO H snjoijdsonj spiodiq SSAllOBJJXa snouaSoijju-uo]^ u snjoqdsoqd I | UOU3EJJ ajqegtuodESun iejoj^ | | I0J3jsa]0H3 I I spiodiq Hsy saAnoEJjxa snouaSojj'iu-uojsj l«£ do 6 b> «  uiajojj I o J3;bav •uj3 aSa am JO iqSia^ f^ f-* b* li^ i^ mvo ID tD lira a s s '^ o -SS ^2 2o.s2 C V C >; Itil t « « c « a c c c c c s c a c c Kc bi be M en 0000 - o tc Uc Vh I- ' >-i 2 2 ^ oC;'S^ o'o.Ji --3 ^:3 2 N T3 Ca N — ^ t-, o f- ^ ^ , ^S^ ^— P- 15 c 3 O O 'E C N li CSS «  H >^ - "I I* ffl>^"0 o 5J C ij!!0.>Sl fc <3 •* o pn 00 n t^ O r^ b>cb b b* op ^00 00 00 M N>0 ri inb'b ■* N OO O l/l w N O " o 00 '-' O' ro O QC 00 30 O 1^00 00 00 00 ■*vO « 00 o o "i-oo o b " 6 N O f^ I 00 N rf C poo PJ 00 M N t^OO OM^ O N 00-^riON^O> f^ iri f*^ r^ f^O m b^ NNNNNNNN COO OO 00 N O' N O 00 r^ o 00 c , N N N\b b CO r*) CM^ Tj- iri sO sO o> N o o t^ 1^ o OO O r^ O^ f^ N f^ ^ b.M «  o ■^oo fn H^ io f^ b* 11 N N N O' 'j- O 30 6 b " ^ N 't ^ in m\0 N (vj f^ inOO V CM MM on in Tj- o ui O o \0 M in ^ r^ 'noo M r^ O'O oo \0 -^ in in r^ ^oo tn tm •III 5 aOl ri^! Omen OU O EI S'^ '^ . -s Hi hcuHHE-' E-h • c o .■5 ^ '-'^^-^ JJ — c^ QQQ C o ^ ^<J8 tfS 0:5 o CO sruoijdsoiij I0J3js3i0q3 spiodiq qsy S3AI}3EJ)X3 snouaSoJi'ra-uoisj ua3oj}Tjsj I I IBJ UI3JOJJ sruoqdsoqj U0U3BJJ spiodiq I I qsy o op saAijOBJixa SnOU330Jl'TU-UO]sI upjOJJ i91EjV\ boob •uj3 333 aqj jo jqSp^ | | moo m N f^ m * •<J-vO O O t~ N o ino \0 o o p p o o o o o b b b b b b b T)-o o^o o o o o ■*p;tr^r»9>J~ b ^ " boovo b t^ t^ £^ t^ r^^ m^o vt 1 I I I I I I I ■~ 1-1 ii « 2 "s ? 3 ^ £3 r r w O .'t! ^ -^ 0, -s^ Q Q m fl^^UKPQ .2S o^'i-ii-a i2 w 0^= •a • E c.E-S 0.1 3 M O C Q 0) u cs 3 0*^ taJ ffiXUJOw IX CI, E-u « .s M _ - ?> C 00 ^ 1 -^ 1 1 1 > 1? °~ ^ T3 C s. N It 'Si 1 rov (188 1903) Conte ( 922) rcmiet & wOO N ,_, o 1^ a "ir^^^f^^Sr^ -^'^i^a Tichoi Farkas Vaney Russo Faur^ E p5 mil s 5 fc^ b o O I 1 A A 000 ° I N 0* N 6 U1 r» 1 o o b b S 5 s s a US «  iapu omct lu 1 &0 cue ••* a-»* a ■s^j: a "c ^"^^ tij c c c '^ ^— ' -000 J3 »■ u u u 3 3 3 ^ !<! « «  i-i CDCfiC/D ^ I 5 a S ^ OS 03 s ^. a ^ •- :2 4) 55 ^ -a ~ ■= <; b J U CO S 2 O •-•2 "55 S 9 E Z 3 312 THE UNFERTILISED EGG AS A [pt. m Fremiet & Dragoiu found the dry material of the egg of the frog to consist of 60 per cent, protein and only 14 per cent. fat. And this has been the experience of all those who have analysed amphibian eggs. Next to the hen's egg, the eggs of fishes have probably received the most attention. The recorded analytical figures for fish eggs are often deceptive, for many analyses have been made of salted fish roes and egg-preparations such as caviar, but the greatest care has been taken not to include in Table 30 results which might have been vitiated in that way. The question is complicated by the fact that analyses of the purified constituent egg-substances prepared from preserved material may well be admitted into consideration, for, except in certain cases, they would probably not undergo much change during the process of preservation. As in the case of reptiles and amphibia, the fish egg is characterised by its predominance of proteins as the food of the growing embryo. It should be remembered that, in all these comparisons, the yolk of the hen's egg is a more proper standard of reference than the whole egg-contents, in which case the differences become even more remarkable. This generalisation appears at all points; thus, the brook-trout, a fresh-water fish, has 30 per cent, of protein and only 9 per cent, of fat, and the herring has 26 per cent, as against 3 per cent. In Table 31 the protein/fat ratios are collected together, and the difference emerges there with great clearness. Though it is obvious that fishes vary considerably among themselves as to the fat-content of their eggs, yet all of them have more protein than fat. The only fishes in Table 31 which come near to being exceptions to this rule are the sturgeon and the dogfish, both of which have an unusually high amount of fat in their eggs (a fact which accounts for the superiority of Russian caviar over other varieties, for the former is made chiefly from the eggs of sturgeon) . Besides this general difference between the egg of the bird and that of the fish, there are many others, but they concern the chemistry of the individual components rather than the rough constitution of the egg as a whole, and will, therefore, be dealt with later on. If Table 31 be further studied, it will be seen that, as far as can be known at present from the few analyses of crustacean and cephalopod eggs, the superior proportion of protein over fat holds good there also. Curiously enough, the only analysis we have for a gastropod egg gives a picture more resembling the egg of the hen, comparatively equal amounts of fat and protein being present. SECT. l] PHYSICO-CHEMICAL SYSTEM 313 The position of affairs may perhaps be summarised by saying that it is only the birds which have been successful in producing an egg really well stocked with fat, though the reptiles clearly show an approximation to this achievement. Does this mean that the storage of fat in the egg is particularly associated with terrestrial embryos? The facts and arguments to be brought forward in later chapters (see Sections 7-7, 9-15 and 1 1-8) make this hypothesis a very likely one, but, as Table 31 shows, the silkworm (the only representative of terrestrial Table 31. Protein/ fat ratio in various eggs. VERTEBRATA Amniota Aves Hen (yolk) „ (whole egg) 0-450 1-035 Reptilia Anamnia Lizard Grass-snake Tortoise 1-450 0-298 2-271 Amphibia Pisces Frog Cod Sturgeon Herring Carp Trout Dogfish Salmon INVERTEBRATA 2-6lO 12-550 1-937 8-333 9-101 3-518 1-050 2-400 ECHINODERMATA Echinoidea Sea-urchin 2-370 MOLLUSCA Cephalopoda Gastropoda Octopus Limpet 5-750 1-245 Arthropoda Crustacea Arachnida Insecta Crab King-crab Silkworm 2-709 3-166 2-139 Annelida Polychaeta Sabellaria 3-162 arthropods) does not seem to have succeeded in storing fat in its egg to any great extent. It might, of course, be argued that this was one of the factors which prevented the insects attaining any considerable size and rivaling reptiles and mammals for the possession of the land. The mammals gave up the heavy fat storage in the egg when they invented viviparity and the fully developed placenta. In 314 THE UNFERTILISED EGG AS A [pt. iii this connection the monotreme egg would be a chemical study of great interest, and it is characteristic of the exasperating fragmentation of this field of work that all we know about the monotreme egg is that its membrane seems to have the properties of a keratin. The suggestion that the metabolism of the fowl, operating on a continuously high level of energy turnover, would naturally tend to fill up the eggs with fat, and is associated with the well-known higher temperature of the avian body (Wetmore) may not be without value, but any special emphasis on fat metabolism in adult birds is precluded by the statements in Schulz's review. It is very significant that as animals became more complicated and more adaptable to varied surroundings, higher, in fact, in the taxonomic scale, they loaded their eggs to a greater extent with yolk. Since the extra material was usually fatty acids, this process appears strikingly in Table 31 . The effects of the yolk have long been familiar to embryologists, and have been best described, perhaps, in a passage by Milnes-Marshall. "The immediate effect of a large amount of yolk", he said, "is to retard mechanically the processes of development, but the ultimate result is to shorten them. This paradox is readily explained. A small egg, such as that of Amphioxus, starts its development rapidly, and in about eighteen hours gives rise to a free-swimming larva, capable of independent existence, with a digestive cavity and a nervous system already formed ; while a large egg such as that of the hen, hampered by the great mass of yolk by which it is distended, has, in the same time, made very little progress. From this time onwards, however, other considerations begin to tell. Amphioxus has been able to make this rapid start owing to its relative freedom from yolk, but now this freedom becomes a retarding influence, for the larva, containing within itself but a very scanty supply of nourishment, must devote much of its energies to hunting for and to digesting, its food, and hence its further development will proceed more slowly. The chick embryo on the other hand has an abundant supply of food in the egg itself and has no occasion, therefore, to spend its time searching for it, but can devote its whole energies to the further stages of its development. Hence, except in the earliest stages, the chick develops more rapidly than Amphioxus and attains its adult form in a much shorter time. The tendency of abundant yolk to lead to shortening or omission of the ancestral history, is well known. The embryo of forms well provided with yolk SECT, i] PHYSICO-CHEMICAL SYSTEM 315 takes short cuts in its development, and jumps from branch to branch of its genealogical tree instead of climbing steadily upwards. Thus the little West Indian frog, Hylodes, produces eggs which contain a larger amount of yolk than those of the ordinary English frog. The young Hylodes is consequently enabled to pass through the tadpole stage before hatching, and to attain the form of the frog before leaving the c:gg\ the tadpole stage is, in fact, only imperfectly recapitulated, the formation of gills, for instance, being entirely omitted." The more yolk, then, the longer the embryo can remain an embryo before having to face the external world, and the more preparations it can make for that event. It is probable that this question is intimately bound up with the penetration of fresh-water surroundings by the originally marine forms. "It has long been noticed", said Milnes-Marshall, following the classical exposition of Sollas, " that marine animals lay small eggs whereas their fresh-water allies lay eggs of much larger size. The eggs of the salmon or trout are much larger than those of the cod or the herring, and the crayfish, though only a quarter the length of the lobster, lays eggs of actually larger size. The larger size of the eggs of the fresh-water forms appears to be dependent on the nature of the environment to which they are exposed. Considering the geological instability of the land as compared with the ocean, there can be no doubt that the fresh-water fauna is, speaking generally, derived from the marine fauna, and the great problem with regard to fresh-water life is to explain why it is that so many groups of animals which flourish abundantly in the sea should have failed to establish themselves in fresh water. Sponges and Coelenterates abound in the sea, but their fresh- water representatives are extremely few in number; Echinoderms are exclusively marine ; there are no fresh-water Cephalopods, no Ascidians, and of the smaller groups of Worms, Molluscs, and Crustacea, there are many that do not occur in fresh water. Direct experiment has shown that in many cases this distribution is not due to the inability of the adult animals to live in fresh water, and the real explanation appears to be that the early larval stages are unable to establish themselves under such conditions. To establish itself in fresh water permanently an animal must either be fixed, or else be strong enough to withstand and make headway against the currents of the streams or rivers it inhabits, for otherwise it will in the long run be swept out to sea, and this condition applies to larval 3i6 THE UNFERTILISED EGG AS A [pt. iii forms equally with adults. The majority of marine invertebrates leave the egg as minute ciliated larvae, which are quite incapable of holding their own in currents of any strength. Hence it is only forms which have got rid of the free-swimming ciliated larval stage, and which leave the egg as organisms of considerable size and strength, that can establish themselves as fresh-water animals. This is effected most readily by the acquisition of yolk — hence the large size of the eggs of fresh-water animals — and is often supplemented by special devices." Here is an explanation for the well-known paucity of eggs in freshwater plankton. In certain cases it is possible to induce an embryo to skip the larval stage which it should normally pass through. Thus Child could abolish the free-swimming larval stage in the ascidian Corella willmeriana, simply by removing the eggs from the parental atrial chamber {p¥L j'^.) to normal sea-water (/>H 8-4). Giard had also noticed the discrepancy in egg-size between closely related marine and fresh-water forms, and had classed it among those cases where like adults have unlike larvae ("Poecilogony"). The classical instance is perhaps that of the shrimp Palaemonetes varians, one variety of which {microgenitor) lives in the sea near Wimereux and has eggs 0-5 mm. diam. (32 1 per female) and another of which {macro genitor) lives in fresh water at Naples and has eggs 1-5 mm. diam. (25 per female). Giard has reviewed this subject in a very interesting paper. "Dans un groupe determine", he said {(Euvres diverses, p. 18), "la condensation embryogenique va en croissant des types marins aux types d'eau douce ou terrestres." The correlated proposition, namely, that the fresh-water forms generally lay fewer eggs than the marine ones, is illustrated by the following instances collected by Carpenter: No. of eggs laid per female per annum A Lamellibranchs ... Gastropods Fishes Crustacea Marine form Ostrea edulis i ,800,000 Buccinum undatum 12,000 Haddock 9,000,000 Lobster 5,000 Fresh-water form Uniopictorum 220,000 Anodonta cygnea 18,000 Average of many snails 100 Average of many limpets 6 Ovoviviparous pond-snails 15 Brook-trout 750 Crayfish 200 Another reason for the poverty of fresh-water fauna was suggested by von Martens who pointed out that the fresh-water climate, with its periods of desiccation and freezing, was much more severe than SECT, i] PHYSICO-CHEMICAL SYSTEM 317 that of the sea. But even these two causes together cannot fully account for the phenomenon, for there are many cases of individual species which they will not cover; thus the Cephalopods, which hatch out as minute but very active copies of their parents, i.e. which pass their larval stage within the egg, and which should therefore be immune from the disadvantage described by Sollas, never penetrated into fresh water. A third reason must be added to those of Sollas and of von Martens. As will be shown in Sections 12 and 13 the marine invertebrate embryo depends largely on the salts of the sea water for its supply of ash, and therefore could not be expected to develop in a medium very poor in inorganic matter. Colonisation of the fresh water could not occur, then, until animals had begun to provide in each egg sufficient ash to make one finished embryo. There seem to be few data concerning the capacity of marine invertebrate eggs to develop in fresh water, although the adult animals have been found often enough to accustom themselves to a fresh-water environment (see the instances given in Semper) . Many studies of the effect of hypotonic solutions on marine embryos can, however, be called to mind, and in all the cases the results are teratogenic. The fate of the Cephalopods, it is interesting to note, is explained by this third factor, for Ranzi has demonstrated the intake of the salts in the sea water by the octopus egg. As for the general statement that animals can afford their young a better chance of survival by providing them with larger amounts of yolk and therefore a longer incubation-period, there is a striking parallel here with the seeds of leguminous plants which are packed with nourishment. In the Origin of Species (6th ed. p. 56), Darwin wrote, "From the strong growth of young plants produced from such seeds as peas and beans when sown in the midst of long grass, it may be suspected that the chief use of the nutriment in the seed is to favour the growth of the seedlings, whilst struggling with other plants growing vigorously all round". It is interesting that the birds show an adaptation exactly similar to the poecilogony of the invertebrates and fishes. Tree-nesting birds are usually nidicolous, but the defenceless state of the newly-hatched squab has brought it about that ground-nesting birds are usually nidifugous. As Table 30 shows, the composition of the eggs of all animals other than those of the frog, the silkworm, and certain fishes, is still, to 3i8 THE UNFERTILISED EGG AS A [pt. m use a phrase of William Harvey's, "hid in obscurity and deep night". It is as yet much too early to try to draw any conclusions from the very fragmentary figures which are all that we have at our disposal, and we may well admit that one of the most urgent needs of chemical embryology is a much wider extension of our knowledge of the static chemistry of the egg. This is a quite indispensable preliminary to the investigation of the metabolism of the embryo in the lesser known forms. The attempt has already once been made to link up in some way the chemistry of the egg with what is known of the type of embryonic development which takes place in it. Wetzel in 1907 analysed the eggs of a sea-urchin, a crab, a cephalopod, and an elasmobranch fish. He pointed out that the eggs he studied were examples of varying richness in yolk, of total and partial, equal and unequal, superficial and discoidal cleavage, as well as chemical systems. Taking the egg of Strongylocentrotus lividus as his first case, he regarded it as typical of a class of alecithic eggs, of a total and equal cleavage type, and he drew attention to the fact that it was rich in water and in salts, but poor in fatty substances, in nitrogen, and in phosphorus. Similarly, in the case of the mollusca, where there is no very definite type of development, the egg of Sepia could not stand as representative of any wider class than the cephalopods, but, as far as it went, it showed that the cephalopod egg was rich in nitrogen, poor in fat and inorganic substances, with a moderate phosphorus and water-content. The decapod Crustacea, to which Maia squinado belongs, have a purely superficial type of cleavage, with no cell-multiplication in that part of the egg which holds the yolk. Accordingly, the egg possessed a moderate fat and water-content, a moderate ash, and much protein and phosphorus. The mammalian ovum is still as unknown chemically as it was when Wetzel was writing, and it may be found to have a constitution not unlike the alecithic echinoderm eggs. For the eggs of birds (and of reptiles, which only differ from them in having very little egg-white) Wetzel found a low protein and water-content, a high proportion of fat and ash, and a large amount of calcium and phosphorus. Here cleavage would only take place at one isolated point on the surface of the mass of food-material. In the amphibia, the richness of yolk, while much more significant than in lower classes, does not reach the level of birds and reptiles. SECT. I] PHYSICO-CHEMICAL SYSTEM 319 and this is duly reflected in the chemical composition by the moderate water-content, the high proportion of protein which is yet only double that of the fat. The case of the dogfish is again different, for there the egg is rich in yolk and the cleavage is meroblastic; thus the water is rather low, the fat rather high, the nitrogen very high, and the ash and phosphorus moderate. But these conclusions of Wetzel's, interesting though they are, cannot really be assessed until a great deal more comparative work has been done. They must rather be taken to represent the kind of correlation we may hope for in the future. However, one of Wetzel's generalisations may be accepted, if with some reserve. He pointed out that the fat-content of eggs showed great variations, rising from 12 per cent, of the dry weight of the Sepia ^gg to 66 per cent, of the dry weight of the (yolk of the) hen's tgg. Again, the nitrogen gave very variable results, rising from 5-3 per cent, of the dry weight in the (yolk of the) hen's ^gg to 6-9 per cent, in the egg of the grass-snake, 1 2 per cent, in the egg of the dogfish, and even in the case of the cod 14 per cent. On the other hand, the phosphoruscontent varied only between the (outside) limits of 2 • i per cent, for the sea-urchin tgg and 3-6 per cent, for that of the grass-snake. Wetzel, therefore, suggested that a distinction might be made, at any rate, roughly, between those constituents of the egg which may serve as sources of energy for the growing embryo, and those which in no circumstances do so. Protein, fat, and carbohydrate would come in the former class; phosphorus (for nucleoprotein) and cholesterol, for example, would come in the latter class. The former would show great variations among eggs of different species, the latter would not. He thus supposed that one might be able to deduce, as it were, the constitution of any given egg, if one knew what substances, and in what proportions, were used by the embryo as combustible material during its development, as well as the constitution of the newly born or hatched organism. From this standpoint Wetzel distinguished four types of substances in the unincubated egg : ( i ) material for the embryo to burn during the course of its development, (2) constituents of the finished protoplasm of the embryo, (3) constituents of the finished embryo, but not for incorporation into the protoplasm itself, but into the paraplasm (in Le Breton's terminology), (4) the protoplasm of the original egg-cell. No aspect of chemical embryology needs attention more 320 THE UNFERTILISED EGG AS A [pt. iii urgently than this, and the correlation of chemical constitution with developmental type should offer a most attractive field for research. But it is not only correlations of this type that lie hidden under the enigmatic character of analytical figures. The water-content of the eggs may have a powerful effect on the sex-ratio, for King found in 191 2 that reducing the water-content of fertilised frog's eggs considerably lowered the proportion of males, while increasing it by means of treatment with dilute acid considerably raised the proportion. A discussion of these facts in relation to genetics as a whole will be found in the review of Huxley. It is probable that the effect which delayed fertilisation has upon the sex-ratio is to be explained by difference in water-content of the eggs. Hertwig was the first to observe this delayed fertilisation phenomenon in some work which he published in 1905, and since then it has many times been observed not only for amphibia but also for trout (Kuschakevitsch; Huxley; Mrsic) . Riddle has suggested that the mammalian egg may be subject to such influences as it passes from ovary to uterus. He quotes van der Stricht's histological work on the bat's egg during this process, and points out that the swelling of the yolk-granules would indicate an absorption of water. The exact degree of hydration of the mammalian egg might thus conceivably have an effect on the mammalian sex-ratio. Table 30 has several more important points which have not, so far, been touched upon. It is interesting to follow in the figures of Milroy the difference between the fish eggs which float at the surface of the water during their development (pelagic ova), and those which sink, or rather float, at lower and denser levels (demersal ova) — the former have a water-content of about 90 per cent., the latter of about 70 per cent. A knowledge of the chemical composition of fish eggs throws a great deal of light upon their distribution in the sea, and so indirectly upon ecological problems. Their fat-content, for example, has been treated from this point of view by Polimanti, whose work will be discussed in the section on the general metabolism of the embryo; and the investigations of the specific gravity of fish eggs, which are discussed in Section 5, have also an important bearing upon these problems. Another point worth notice is the approximately constant percentage of cholesterol in different eggs, nearly always about 500 mgm. per cent, of the wet weight, a proportion which, roughly speaking, holds for the egg of the hen as well. SECT, i] PHYSICO-CHEMICAL SYSTEM 321 It would be as well to emphasise the fact that no principle of selection has been used in the preparation of Table 30, on the ground that results such as those of Roffo & Correa on a Brazilian gastropod, and McCrudden on fresh-water fishes, which seem obviously wrong, may not be so at all. The estimation methods and analytical processes which are by general consent judged most satisfactory at the present time cannot be considered in any way final, and to have excluded certain results on account of the technique employed in obtaining them would not have been justifiable. Table 30 does not, therefore, absolve investigators fi'om the duty of looking up the original papers in such cases as touch them most closely, and forming an independent judgment, according to the best opinion of the time, on the stress which can be laid upon them. It is needless to say that I leave out of account all doubtful figures in the generalisations made here. I -12. Egg-shells and Egg-membranes Very little is known about the relative proportions of yolk, white, and shell, in the eggs of the lower animals, or rather, in most cases, egg-contents and shell or surrounding membrane. Table 32 gives a few figures. The discrepancy between the results of Ford & Thorpe, on the one hand, and Wetzel, on the other, is very strange, especially as they both used Scyllium canicula eggs, but it is probably due to insufficiency of the statistical element. Ford & Thorpe's proportions are more likely to be accurate. Much work, however, has been done on the membranes and hard coverings which invest the unincubated eggs of diflferent kinds of animals. For instance, the gelatinous substance which surrounds the undeveloped amphibian egg was examined chemically by Brande in 1 810, who noticed that it absorbed water and was not precipitated by tannin or by strong acids. Later work has shown that it consists almost entirely of mucoprotein and water. Wetzel's figures for its weight are shown in Table 32. Giacosa isolated mucin in a pure state from it in 1882, and the figures which he obtained for its percentage composition are shown in Table 33. He was able to show the presence of a reducing sugar on hydrolysis, but he could isolate nothing else from the jelly, and therefore concluded that it was pure mucin. The presence of glucosamine in the mucoprotein was afterwards confirmed by Hammarsten, by Schulz & Ditthorn and by Wolfenden, who confirmed Giacosa's finding that 322 THE UNFERTILISED EGG AS A [PT. Ill Table 32. In % of total egg-weight Species Egg-membranes Whit( Herring Carp ... Cod ... Pike ... 2-4 3-7 4-4 4-1 — Dogfish Silkworm 5-4 26-9 8-87 (wet) 19-3 36-5 Trout ... Octopus 25-97 (dry) I3-57-20-29 86-0 — Yolk Investigator and date — Konig & Grossfeld (191 3) 3? 33 75-3 Ford & Thorpe (1920) 36-5 Wetzel (1907) — Tichomirov (1882) 14-0 3J Kronfeld & Scheminzki (1926) Ranzi (1930) Tomita's figures. Marine turtle ( Thalassochelys cortica) Weight Shell White Yolk Total m gm. 2-0 13-5 18-9 34-4 % 5-8 39-2 55-0 Wetzel's figures. I Frog {Rana temporaria) Ovarial egg (no jelly)... Egg with unswoUen jelly Jelly alone Swollen jelly ... Water content of ovarial egg jelly „ „ egg and jelly Empty dry jelly Dry egg Dry egg + dry jelly ... Thus of dry weight egg jelly Weight in mg. 1-897 4-674 2-777 8-97 0-62 0-90 1-52 Melvin's figures. Shell-weights of insects Squash-bug {Anasa tristis) Luna moth [Tropoeoa luna) Cecropia moth {Sarnia cecropia) ... Smartweed-borer {Pyrausta ainsleii) 52-5 78-65 67-48 59-28 40-72 % of total weight of eggs 29-2 23-3 22-0 31-0 it was remarkably resistant to putrefaction, and studied the effect of enzymes such as pepsin upon it. The resistance of frog ovomucin to putrefaction was for long a puzzle to biochemists, but it seems to be explained by the unwillingness of most SECT, i] PHYSICO-CHEMICAL SYSTEM 323 bacteria to grow on pure proteins, and as the jelly contains no enzymes of an autolytic character no protein breakdown products are formed, and consequently no bacterial growth takes place. This might be considered a protection of the developing embryo from bacterial attack. It is very probable, moreover, that the mucoprotein acts as a source of nourishment for the young tadpoles immediately after hatching, for they invariably attach themselves to it after they emerge from the egg-membrane, and hang on to it by their oral suckers (for histological details consult Nussbaum and Lebrun). On the other hand, development will readily proceed in the absence of the jelly, for as Hluchovski has shown it is disintegrated by exposure to ultra-violet light and may thus be removed without harming the eggs. The swelling which takes place in the gelatinous covering when the eggs are shed into the water was studied as long ago as 1824 by Prevost & Dumas, who measured the size of the eggs at intervals after they were laid. Their table is as follows : Hours after laying Diameter of egg (mm.) 2-5 1-5 5-0 2-5 6-3 3-5 7-1 4-5 7-2 5-5 7-1 6-5 7-3 They observed that dyes would pass through the jelly as soon as it had swollen, but not before. Similar work by Wintrebert on Discoglossus pinctus gave the following figures : after laying Diameter of egg (mm, o-oo 2-5 X 2-3 0-03 3-0 X 2-7 016 3-3 X 3-0 0-66 5-8 X 3-2 800 5-OX4-6 As regards the mineralogical and morphological structure of the egg-shells of the lower animals, a good deal is known, and for full detail the reviews of Prenant and of Biedermann should be referred to. The majority of reptile egg-shells have their calcium carbonate in the form of calcite, as Kelly; Schmidtt, and Meigen have shown, but the two first-named investigators discovered that the tggsheUs of chelonia were of aragonite, and later Lacroix observed a similar phenomenon in the case of certain saurians. The tgg -a a »-l ^-v a plH lO u o d V5 ,— ,0^ .§5 V cco_i B > '0 cS "C a S fe S E -= '*;i: 3 cS r3 k! u >ffio;z: |i| CO p^ i^ ^ ?, - o SOOS o bp '2 co^ cr ■" ^ fi i- -> S w JJ p U . S V! C3 O O^ O Cj :0 Ui ^ Si S2 c .E > ffi T3 ^, ^ < oac. ca V V - > > u u V hJi-I u ■^ (S CO f~. i^ r^ CD ^ 6 Th' lO m 1 1 1 ^ 1 1 1 ^ LO O) ■^ 1 1 1 1 Tj- 1 1 1 ^ 01 CO S (N <N (M tr> CO "^ -co locf J^ •^ CO CO 6 ' ' '-' 066 i-l Ci 6 V 1 CO (M y 01 CO cog CO coo <>' CO r^CT^T** in CO «  6 -H 6 « 6 6 6 T^COCOw CO ix> ■* 0^^ «3 CO to CO inOi ^ "^ lO lO io 10 tJ" N4 »-t 0^ 1 1 r^ 1 m loco 1 1 CO ^90 CO 1 1 f,.r- 1 ' ' f^f- f^ r^!0 1 1 ^"^ 1 1 1 " ^^ CO to 1 1 04 CO 1 \ 1 COM CO lO lO 10 Tf" CO II I ^ COtO COCO " lOTfOO," r^ lO lO r- f^ O CO '^'-o <-0 O O " o o O " o o 6 "^ lO lo rhco r~- 01 coto COCO " coco O) (O 10 in loto CO I oy^irS^l 2L^~ a)iocr)"-*o< I o^inpr^l ^'TP o«o»-'"" i,M«i-<i-i 000 inoj 00050 -^o o^5< ;*co eo eo inioinininin'OiO'^eo o o •* Oi r-~ o CO CO (O coco <o r^to m I T^jH in CO I 01 0) 01 irj eo CO S- - ^ lij i^ e.9 •2t3 3^ _c j^xi a, 3 o 3 -Cl -fi -TJ cj 3 S -s y .^-S 3 t

-yS .SJ^ fc. Q. lU .. peovin akdown

hthulin: Paranucl Ichthulic .5 ■5 S "-S 3 <u.y .S Ui u 1— 1 G«  u P ^-5 3 (I 3 lU 3 T3 3 p c« 

§ a o.g .0 cs c'3 g3'3^-" it^.-s^

dono us b <i "^ r^ « 

-S^-| i: c 1 7. :: m « iS < r. u ^ ►- 2 TILl se sna iile CO s Rep Tortoi Grass Croco( Q. — ' < 2 fa -'-'i b ^ s . . S ,« • •"w S^ M 3 0, fcj HOUUffi • 3 «  =«  Co • W)--^ bo c « S - ^ •"CTs-i CT3 t^TJ >>u J; S « t, O^ O t« c« u e) u 11 O tJ< CO CO 01 " c^ c< C be rt ^<2 c3 « O 3 3 g 3 Ih Ci< ^ ^co .^^ 01 CI ^ OICO o w c 2 M O o " u o >S9 ►" t< " ea o w O bo JS CO '-' +3 ^ ^ (-, bD 3 o C t; O lO 10 060 O) o 0» Tt" o»<r> o o CO lO x^ ■*co o O c< ij* CI ^ U 0< CO" t~-T}<T}>C5-* C^OOO-OOCOtJ" ^66666666 0'-io»oo»cooco O ■<t' C< coco '-' t^'O C5 lO O coco 1^ I I I I CT> •* LO COK3 CD ii CTlCO >pcp O ~ CO | | | | ^ O o r^ CT) inco CO CO ■* -" lO •* ci tOl^CfOO'-iOCOOlCOI^t^'-'O in o< T^ Lo^i in lo ■^ -^ LO'-is -rt" ot tJ" 0<tJ<O0<OOOCD coto r^Loo<^cp o< io io io io LOCO " t} (V f^ C50 " CO p o( o I "4f ^^ o< I ot I r^co I ^ 'p CO'^ o UD i^ r^ •- lO LO ^ CO - CD - CD CO CDco cr> r^ co'O ■* 10 ■-■ LO CO LO t^ f^ f~(i) to o o o CJ5 o» to CT> '*'-^ LO incp to CO lO 0) CO" CO" OI^lOn «  lO'OLOtOiOiO'^'^iOiO <o - - ei C^i-D CO CO t-D f^to to rt :; ;i 3'3'S'3 J2JS "3-C U U - tJ K>l C C fl c c be— ■?!— d

^ "YolkIchthi ThuicI Ichthu Kerati hthu bum hthu M ^J2 ^<^ <^ nS

" — — — G*"*-* C — 3 3 3 3;S'3"3-S 3 „^^^-c;5x-c (S-3 cjooo'jaos/t-> bO bo « =s S g 'I O 1 : :s2 ? ? - : • * C I 1^ :2.S ,0 -2 ^ "g a c3 ^ g^'C ^r*^ ^ -_- :cr)a,05i>5 ^' « '. iS ^ C •^--' bo c bo "3 u bo ' he ' .S c «  C &,'t^ 3 ■ 2 t«  C "0 a; i- I-c IJ (- > w cs ii H(^OQ X ho< ffiUEH u bo Ki O OQ anjs — c 'op a o z ^ ^3 2, 3 c , a' ^ ^ 6^ 326 THE UNFERTILISED EGG AS A [pt. m membranes of snake's eggs which show all variations as to lime-content (see Table 9) are, as Kelly has shown, composed of amorphous and unstable calcium carbonate. The eggs of gastropods, such as Helix, Ampullaria, Bulimus, Amphidromus, etc., are, as Turpin {Helix aspersa) and Rose {Helix pomatia) , besides the workers mentioned above, have demonstrated, like those of birds in having their lime in the form of calcite. For a general theory explaining these differences see the paper of Prenant. The shells of eggs may also contain calcium phosphate. In the hen and in birds generally there is very little, but the globules seen in their egg-shells are believed to be calcium phosphate, though no analysis has given a figure of more than i per cent, of this salt. In other eggs, however, there may be more; thus Gmelin found 7-3 per cent, in the egg-shells of a tortoise, and Kelly noted its presence also in those of Bulimus and Lophohelia, though she gives no analytical figures. It is interesting to note that the mineralogical form of lime in the egg-shell may vary during the development of the embryo; thus Kelly says that the shell of many full-grown mollusca is conchite, while that of their respective embryos and eggs is calcite. Kelly found that the organic substance was a remarkably constant proportion of the shells of mollusca, reptilia and birds (see Table 9). Some eggcoverings contain almost no water at all (birds), others have more than the egg-contents, as has been shown for the trout's egg by Kronfeld & Scheminzki (membrane 75 per cent., egg 66 per cent.). By far the commonest substance of which egg-membranes are composed is keratin, though this protein seems to take many forms, and not to have exactly the same properties in different situations. The earlier workers were content to assert the presence of it on the basis merely of solubility tests. Thus in 1874 Schenk studied the egg-shell of Raia quadrimaculata, and decided that it was 95 per cent, keratin after the application to it of the protein colour reactions and an examination of its behaviour towards various solvents. The same conclusion was arrived at by the same methods by Hussakov & Welker for the egg-cases of Raia erinacea, and the Port Jackson shark, Heterodontus philippi. The keratin of these egg-cases was insoluble in all solvents except acid and alkali. They found that sulphur was present, but no phosphorus, and they were unable to find any reducing sugar after total hydrolysis. Irvine, using an optical test for chitin, found SECT, i] PHYSICO-CHEMICAL SYSTEM 327 none in elasmobranch egg-cases. Krukenberg in 1885 decided that the egg-case of Scyllium stellare was of a keratinoid nature, because of its percentage composition, in which he found a marked amount of sulphur. He observed the interesting fact that the egg-cases of this fish, while still in the uterus of the parent animal, would dissolve in pepsin and trypsin, while after they were laid they would not dissolve in solutions of either enzyme. He also isolated tyrosine and leucine firom the keratin of the egg-cases of Scyllium stellare. He made very similar researches on the egg-cases of Scyllium canicula and Myliobatis aquila, finding that they possessed rather different properties and seemed to be of different constitution; thus on hydrolysis he recovered a great deal of leucine and hardly any tyrosine from the keratin of Scyllium canicula, while from the keratin of Myliobatis the yields were precisely reversed. The latter substance was also considerably more resistant to digestion than the former, and Krukenberg considered that the former was not a keratin at all. He had already decided (wrongly, as it turned out) that the shell-membrane of the hen's egg was mucin, not keratin, and now he concluded that this also applied to the egg-case oi Scyllium stellare, as well as to that ofLoligo vulgaris, of which he made a separate examination. He thought it possible also that the jelly which surrounds the egg in the ovo viviparous selachians might be a mucin too, especially as, according to Schenk, it was not precipitated by chromic acid, and he himself found that it was extremely resistant to digestion by enzymes. This material has received no further chemical investigation since the time of Krukenberg. Other workers who identified the proteins of egg-membranes by the aid of colour tests and solubility reactions were Leuckart, who showed, as far as anything could be shown with such preliminary methods, that the membranes of planarian eggs were of chitin, and Yoshida & Takano and Jammes & Martin, who drew a similar conclusion about the coats of the eggs of Ascaris lumbricoides, which they found were readily soluble in gastric juice or in any acid.^ The case of the parasitic nematodes is of special interest, for the chitinous membrane does not arise until after the fertilisation of the egg, being, therefore, in a sense, analogous to the fertilisation membranes of echinoderms. Whether the chitin is formed as it is required during these early stages, or whether it is already present in the unfertilised egg-cell in some soluble form, is uncertain. Faure-Fremiet in an ^ See also Campbell on the chitin of insect egg-membranes. 328 THE UNFERTILISED EGG AS A [pt. iii attempt to throw light on this question, prepared pure samples of chitin from the newly fertilised eggs ofAscaris megalocephala by boiling them with strong potash, and identified the chitin chemically, isolating glucosamine hydrochloride from it. Remembering that Weinland showed that chitin is probably formed from glycogen during insect metamorphosis, Faure-Fremiet estimated the glycogen in the Ascaris eggs before and after fertilisation. Before fertilisation there was an average amount of 20 gm. per cent, dry weight, but afterwards only 4-67, the extreme values being 5-91 and 3-23, so that no less than 17 per cent, of glycogen had disappeared. Estimations of chitin in the egg-envelopes after fertilisation gave results of between 8-3 and 10-7 per cent, dry weight of glucosamine (calculated as glycogen) with an average of 9-23. The total glucose, then, in the fertilised eggs was 12-83 to 15-08, as against 20-0 in the unfertilised ones, a loss of 7 to 9 per cent. All the glucose lost, therefore, could not have transformed itself into chitin, but must have had some other destination, perhaps butyric and valerianic acid if Weinland's view is correct. The eggs o^ Ascaris have also an " ovospermatic membrane", but for the discussion of the significance of this reference should be made to the memoir of Faure-Fremiet, and nothing is known about it chemically. Their third membrane, the internal one, would seem to be composed to a large extent of ascaristerol (see p. 352), for the histological evidence demonstrates a collection of the ascaristerol globules at the periphery of the cytoplasm. After fertilisation, FaureFremiet found the saponification number of ascaristerol lowered from 199 to 145, from which he concluded that its constitution had been slightly altered. Zavadovski has also described the egg-shells of many nematodes. • Neumeister, who found more than 5 per cent, of sulphur in the shells of the reptiles, Calotes jubatus, Ptychozoon homalocephalus, and Crocodilus biporcatus, concluded that they consisted of a true keratin, and the reactions given by the egg-membrane protein of a monotreme, Echidna aculeata, led him to the same conclusion in that case also. Table 9 gives the figures which he obtained for the calcium and other constituents of some of these egg-shells, as well as the very similar investigations of Wicke & Brummerstadt on Alligator sclerops. From these fragmentary results, it would seem that the eggmembrane protein is here keratin, and a quantity of calcium is secreted into the membrane by the animal, varying in amount from SECT, i] PHYSICO-CHEMICAL SYSTEM 329 90 per cent, to 10 per cent,, according to the species. Again, the egg-membrane of the Brazihan gastropod studied by RofFo & Correa is said, on the basis of qualitative tests only, to be a true keratin, containing no reducing sugar and associated with no other substances, save 2*45 per cent, of ash. It contained calcium the amount of which did not vary during development. The transparent horny egg-membrane of the selachian Mustelus ^ laevis, which disappears half-way through the development of the ■ embryo, has also been investigated by Krukenberg, who compared it with the egg-membrane of the grass-snake, Tropidonotus natrix. The former resembled the shell-membrane of the hen's egg rather than the true keratin of the Myliobatis egg-case. The latter seemed to have some of the properties of elastin and some of those of keratin ; from it he was able to isolate a reducing carbohydrate as well as glycine, tyrosine and leucine. Krukenberg was also one of the earliest workers to make quantitative investigations on this subject. His figures for the protein of the egg-shells of Murex trunculatus and the whelk Buccinum undatum, which are given in Table 33, led him to make a new class of such substances, the conchiolins. As no data exist for the sulphur content of most of these proteins, it is impossible to say whether they are keratins or not, and the whole subject needs re-investigation. About five years later, Engel also investigated the egg-membrane protein of Murex, and, obtaining 0-5 per cent, of sulphur from it, concluded, its other properties taken into account, that it was a keratin. Engel also agreed with Hilger, whose figures for the egg-membrane of the snake, Coluber natrix (see Table 30), suggested an elastin as its principal component. He had not been able to find any sulphur in it. About the same time, Wetzel examined the conchiolin in_the_egg-shells of, Mytilus edulis, and obtained from it, after hydrolysis, leucine, tyrosine, glycine, Various hexone bases and ammonia, but no phenylalanine. The first efforts at quantitative discrimination between egg-membrane proteins were contented with ascertaining the elementary composition ; thus von Fiirth analysed the protein of Loligo vulgaris eggs in this way (39 per cent, glucosamine), and Verson, and later Tichomirov, decided that the egg-shell of the silkworm, Bombyx mori, was a keratin-like body (3-7 per cent, of sulphur), though, owing to its unusual properties, they called it chorionin. Of these two lastnamed analyses, it is probable that Tichomirov's is the more accurate. 330 THE UNFERTILISED EGG AS A [pt. m for he was more careful to remove all the adhering silk than was Verson, and Farkas' independent work agrees rather with his. It is, at any rate, clear that the shell-substance of the silkworm's egg is not chitin. According to Lavini the inorganic constituents of the silkworm egg-shell are potassium silicate, sulphate, and carbonate, to the exclusion of all other salts. The work of Pregl and of Buchtala in 1908 is perhaps the most thorough investigation of the amino-acid distribution of an eggmembrane protein. The figures they obtained are given in Table 34. The keratin of the egg-case' of Scyllium stellar e was the only one of which they made a complete amino-acid analysis ; for that of Pristiurus melanostoma and Scyllium canicula they only determined the cystine content and large groups such as the monoamino-acid nitrogen. Scyllium ovokeratin seemed to follow very closely in its constitution the ovokeratin of the hen, according to the figures of Abderhalden & Ebstein, which have already been discussed, but separated itself off very sharply from it on account of its high tyrosine content. The ovokeratin of the tortoise Testudo graeca, which had been investigated two years previously by Abderhalden & Strauss, was again different, having no tyrosine, but a very high percentage of proline. As far as this work goes, it would seem right to conclude that, though the eggs of different species may use similar proteins in their external membranes, the constitution of these proteins may vary very considerably. The work of Steudel & Osato, and of Osato, however, brought a new factor into the problem. Their analyses of the egg-membrane protein of the herring's egg, which are shown in Tables 34, 38 and 39, gave results which differed from the usual keratin figures, but which very closely approached the analyses which they were making at the same time of the ichthulin of the herring's egg. Thus the amide nitrogen (2-05 per cent.) was lower than any of the keratins, but approximated instead to the i-8i per cent, of herring ichthulin. What appeared to be the case on a general survey turned out to be certainly so when the amino-acid distribution was examined, for the two sets of figures almost exactly corresponded. The properties of the eggmembrane protein and the minute amount of sulphur in it precluded its classification as a keratin, and the fact that no reducing sugar could be discovered among its breakdown products was convincing evidence against its being a mucin. Osato suggested that it was 330* Table 34. Distribution of amino-acids in egg-proteins. Amide N ... Glycine Alanine Valine Leucine Proline Phenylalanine Aspartic acid Glutamic add Serine Tyrosine Cystme ... Hiiiudinc ... Arginine Lysine ... Tryptophane Humin Unidentified Total di-amino Total mono-am Non-amino N c ii Scl U tl Present 19'40 — 53-73 2-76 4.4 _ ^ ^Jolie O'bo 103 10-6 021 0'44 0-30 i-H'i 114 '•7 0-32 '3'5l' l-o6 3-2 '■45 io-i6 0-29 3-7 — — 9-B6 — None Present Present Present Present Present 50-7 — 2-56 9-04 Present 13-72 3-8o 0-37 4-i6 — None 6 3-8 3-66 Ii-ig o-t 0-93 0-62 0-34 None 82 8-83 Present 'None 92 n-og o-ig O'SS 07 8-33 0-39 0-22 l-8l 2-05 7-56 — ■■77 — — — 6 1 -55 3.90 None SECT, i] PHYSICO-CHEMICAL SYSTEM 331 simply an insoluble modification of ichthulin. As he pointed out, industrial use has long been made of insoluble forms of proteins, such as casein, and there was no reason why the egg-membranes of certain eggs, at any rate, should not be insoluble modifications of the proteins of their yolks. Steudel & Osato also suggested that the ovomucoid of the egg-white of the hen might be a phylogenetic reminiscence of the mucoprotein with which the amphibian egg is surrounded. For a review of this work see Steudel. The eggs of salps and tunicates are surrounded by a coat of very much smaller cells which act as some sort of protection for the developing embryo inside. Zavattari has demonstrated histochemically the presence of an abundance of glycogen in these test cells, and believes that they have a nutritive function. If so, this would be a third case where such an active participation of the shell or case in embryonic metabolism would have been noted, the two others being the abstraction of calcium from the shell of the hen's egg, and the contribution of amino-acids by the egg-case of the silkworm. A good deal is known about the osmotic and other properties of the membranes of amphibian and fish eggs, but these are so intimately associated with the physico-chemical processes taking place during development that consideration of them will be postponed to Section 5. It will suffice to mention here the experiments of Peyrega, who found that the egg-cases of Scyllium canicula were permeable to salt. He fitted up osmometers with small pieces of the case as the membranes, and observed that it took about 20 days to establish osmotic equilibrium with respect to solutions of sodium chloride about as strong as sea water, when distilled water was put on the other side. These egg-cases have also been shown by Needham & Needham to be permeable to urea and ammonia. 1-13. Proteins and other Nitrogenous Compounds The principal protein substance which is found to occur in the eggs of all known animals closely resembles the vitellin of the hen's egg. It has even been found, according to Chatton, Parat & Lvov, in the food-reserves of infusoria. The early analyses of the eggs of the pike by Vauquelin in 181 7, of the barbel {Cyprinus barbus) by Dulong d'Astafort in 1827, and of the trout [Salmo fario and Cyprinus carpio) by Morin in 1823, ^^^ to no more than the view that an albuminous substance w£is present in them. But with the work of Gobley on the hen's egg, which has already been described, a more solid basis for comparison was achieved, and Valenciennes & Fremy, in a memoir which received a prize from the Academy of Sciences and which was translated into English, proceeded to examine the eggs of as many species as were available to them. Gobley's only excursion into comparative chemical embryology had been a detailed analysis of the carp's egg, published in 1850, but he had not been slow to point out the differences between this analysis and that of the hen's egg. His figures are shown in Tables 2, 30 and 33, where it will be seen that he got a value of 15-76 per cent, protein (wet weight) for the hen, and 14-23 per cent, for the carp, but 31-43 per cent, fat for the hen and only 2-57 per cent, fat for the carp. The carp's egg had, he found, about 10 per cent, more water than the yolk of the hen's egg, but only a third of the lipoid substances. Fremy & Valenciennes specially directed their attention to the protein fraction, and attempted to discover whether the vitellin was the same in all eggs. For the most part they relied on histological appearances (the "dotterplattchen" were greatly discussed at this time), but they also examined the solubility relationships of the proteins from each egg, and in some cases subjected the purified substances to elementary analysis. The figures they obtained for the different compounds are all given in Table 33, and the eggs they investigated in Table 35. They were able to isolate a number of vitellin-like proteins, soluble in salt solution and precipitated by the addition of water. They compared vitellin with fibrin, and concluded that the two substances were almost identical, in spite of slight differences in the analytical figures — "for bodies of this nature", they said, "which are not crystallisable and insoluble in water and which are therefore very difficult to purify, where is the chemist who could answer for i per cent, of nitrogen in an elementary organic analysis?" Ichthin, which they isolated from fish eggs, differed from vitellin by not becoming an opaque mass when placed for a long time in boiling water, and by giving a violet instead of a blue colour when treated with boiling hydrochloric acid. Ichthidin, another product offish eggs, differed from ichthin in being soluble in water. Ichthulin, the third member of the group, differed from the others in not being soluble in all dilutions of saline, but in being precipitated from the aqueous extract by further addition of water. As for emydin, it closely resembled ichthin, and it is SECT. l] PHYSICO-CHEMICAL SYSTEM 333 not easy to see why Valenciennes & Fremy did not identify it with that substance. The remaining egg-proteins, which they did not further investigate, they referred to under the generic name of albumen. Table 35. Investigations AvES Callus domesticus Pisces of Valenciennes & Fremy. Vitellin Elasmobranchs Raia clavata Ichthin Torpedo martnorata >j Scyllium canicula 99 Galeus canis jj Alustelus laevis 99 Squatina angelus 99 Raia fullonica >j Raia rubus j> Teleosteans Cyprinus carpio Ichthidin and ichthulin Labrax lupus Ichthulin and ichthidin Alugil chelo Scomber scombrus Pleuronectes maximus Pleuromctes solea Solea armorica Unidentified species of salmon >) eel Albumen Reptilia Testudo mauritanica Emydin Cistudo europaea jj Unidentified species of lizard Vitellin jj grass-snake )> >> viper „ (?) Amphibia }> frog Ichthin >> newt 5> Crustacea jj lobster Albumen Ar-achnida and Insecta — — Albumen MOLLUSCA — Not albumen The differences between the compositions which Valenciennes & Fremy found for these substances are not great, and it is very doubtful whether they are more than modifications of the same substance, especially as these workers admittedly had great difficulty in obtaining pure preparations. But the problem of the identity of the vitellins is not yet settled. The later investigations are all grouped together in Table 33, and the differences between the preparations can easily be seen to be small. The work of Plimmer & Scott proved that ichthulin is a phosphoprotein closely allied to vitellin. Among the more interesting observations must be mentioned those of Levene & Mandel; Levene, and Walther, on ichthulic acid obtained from the ichthulins of various fish eggs by digestion 334 THE UNFERTILISED EGG AS A [pt. iii with pepsin and other methods. These with their very high phosphorus content approach closely the " paranucleins " or vitellic acids obtained from the vitellin of the hen's yolk by Levene & Alsberg and others. Evidently there are several possible stages of breakdown, for Walther's ichthulic acid only contains 2-8 per cent, of phosphorus, while that of Levene has as much as 10-4. Here, also, however, there are great variations; thus, while nearly all the ichthulins studied have from o-6 to i -9 per cent, of phosphorus, the preparation of Steudel & Takahashi from the herring's egg has only 0-014 P^r cent. In the yolk of a dogfish egg, Zdarek found no less than three proteins, the third of which may possibly correspond with Konig & Grossfeld's albumen class. In 1908 Alsberg & Clark claimed that phosphorus was quite absent from the principal protein of the egg of an ovoviviparous selachian, Squalus acanthias, but some twenty years later I re-examined the question and obtained without difficulty o-6 per cent, from selachian ichthulin (derived from the same species). This yolk also contains a second protein, thuichthin, corresponding closely in properties and constitution with the ovolivetin of the hen studied by Kay & Marshall (see Tables 10 a and 33). Gray has studied the properties of the ovoglobulin or ichthulin of Salmo fario. If the yolks are poured into water, a dense white clot is formed and the water becomes cloudy. The precipitate is soluble, however, in acids, alkalies and neutral salts. When the egg-cell dies, the egg becomes opaque, and this must certainly be due to the precipitation of the globulin, for by placing dead white eggs in normal sodium chloride solution they rapidly become clear and resemble normal eggs, but regain their opacity when removed to distilled water. The clearing process takes 15 minutes but the precipitation takes i| hours. Evidently the dead protoplasmic membrane can no longer retain in the egg the electrolytes necessary for solution of the ichthulin. Further work on the properties of teleostean ichthulin was done by Runnstrom. 1-5 parts of egg "Pressaft" having been added to 1-28 parts of water and the ichthulin precipitated, the effect of various ions on its solubility was tried. The anions placed themselves in the order: SON > I > NO, > SO. > CI > acetate. SECT. I] PHYSICO-CHEMICAL SYSTEM 335 Thus for 2 c.c. of potassium chloride solution, 0-3 c.c. of distilled water had to be added to get coagulation, but to 2 c.c. of KSCN solution, as much as 6-4 c.c. The cations went as follows: Ca > Mg > Sr > K and Na. The egg-white of the dogfish egg was thought by Brande in 1810 to be identical with the jelly surrounding the egg of the frog, but whether the former really consists of mucin and not albumen cannot be definitely stated, for no work has since been done on it. However, my wife and I, in our work on the eggs oi Scyllium canicula, frequently observed a coagulation of the egg-white with acetic acid, which would point to the latter possibility. The proteins of the echinoderm egg have never been properly investigated. Vies, Achard & Prikelmaier have estimated from cataphoresis experiments that the average isoelectric point of the Paracentrotus lividus egg-proteins lies between 5-0 and 5-8 pH., but their grounds for this figure are not free from criticism. Vies & Gex, in some interesting experiments, have studied the normal unfertilised sea-urchin's tgg spectrophotometrically. The absorption spectrum of the normal egg has peaks or bands at wave-lengths of 490, 395, 370, 315, and 230 Angstrom units, and a marked trough between 260 and 240 A. This curve is very peculiar, for on the one hand it shows much transparency in the ultra-violet although most organic substances do not, while on the other hand there is nothing at all corresponding to the bands of absorption about A 275 which all proteins give. This absorption is brought about by the cyclic amino-acids in the protein molecule, and it is quite impossible that these should be altogether absent from the egg-proteins of the sea-urchin. Vies & Gex considered various technical possibilities which might explain these effects, but did not think that any of them would account for what was perhaps the most remarkable part of the investigation, namely, the finding that on cytolysis ("white") a perfectly definite and clear absorption spectrum for protein revealed itself In the intact egg, then, this must be masked by something else. Speculation on the nature of this mechanism would be easy, for all kinds of eflfects might be responsible, e.g., formation of complexes, reduction equilibria, and satisfaction in vivo but not in vitro of residual valencies in the protein molecule. If this very interesting work should lead in the future 336 THE UNFERTILISED EGG AS A [pt. iii to a revivification in a subtler form of the old biogen molecule theory (though it is to be hoped that it will not), not only as regards the egg-cell but as regards protoplasm in general, we shall at any rate possess in the spectrophotometer a powerful means of studying the untouched normal cell-interior. Doubt exists with respect to the presence of reducing carbohydrate in the ichthulin molecule. Levene & Mandel obtained minimal quantities of laevulinic acid from their cod ichthulin, but this finding was associated with the presence of purine bases. Six years earlier Levene had been unable to find a trace of glucosamine in cod ichthulin. Similar negative results were obtained by Steudel & Takahashi on herring, and by Hammarsten on perch, ichthulin. But the presence of glucosamine in notable amounts has been reported for Torpedo ichthulin by Rothera, and for carp ichthulin by Walther. While it is possible, and even probable, that ichthulins from different fish eggs may vary much, it would be very desirable to know to what extent this is the case, and a comparative study of ichthulins is much needed. As we have seen Levene & Mori have isolated a trisaccharide from avian vitellin. Closely allied to the question of the presence of carbohydrate groupings in the ichthulin molecule is the equally disputed problem of the presence of purine bases in the undeveloped tgg. We have already seen that Miescher's identification of nucleoprotein with vitellin was quite erroneous, and have described how he was set right by Kossel. For the hen's &gg, it is now fairly clear that nucleins are present only in exceedingly small amounts at the beginning of development, not exceeding, for instance, i or 2 per cent, of the total nitrogen or phosphorus. But there has been more difficulty in deciding what is the real state of affairs in the eggs of fishes and aquatic invertebrates. Walther (carp), Hugounenq (herring), Linnert (sturgeon), and Hammarsten (perch), all examined the ichthulin of these eggs for nucleic acid, and all failed to find the least trace of it. Henze, on the other hand, working with the whole tgg of the cephalopod. Sepia officinalis, isolated considerable amounts of purines together with no less than 1-15 gm. per cent, of a pentose. Tschernorutzki a little later found that 10 per cent, of the total phosphorus of the herring's egg could be accounted for as nucleoprotein phosphorus, and the nucleoprotein itself amounted to i-ig gm. per cent, dry weight. Masing; Tichomirov, and Needham & Needham reported SECT, i] PHYSICO-CHEMICAL SYSTEM 337 quite similar results with the sea-urchin's egg, the egg of the silkworm and the eggs of various Crustacea, echinoderms and an annelid. In the sea-urchin egg purine bases were found accounting for 6 per cent, of the total nitrogen as nucleoprotein nitrogen, while in the case of Bombyx there were 20 mgm. per cent, dry weight. Again, Levene & Mandel isolated from their ichthulic acid in 1907 0-344 P^^^ cent, of guanine, 0-307 per cent, of adenine, 0-360 per cent, of uracil and 0-309 per cent, of thymine. Mandel & Levene were also able to isolate nucleic acid from cod's eggs. It would certainly appear from this evidence as if ichthulin and vitellin may be associated with small quantities of nucleic acid. In this connection it is of interest that Calvery has evidence that the chick embryo can synthesise "yeast-" as well as animal nucleic acid. Steudel & Osato have also obtained guanine and adenine from herring's eggs, but this was in the nonprotein nitrogen fraction, and there was therefore no evidence from their work that any preformed nucleic acid was a constituent of the egg. The most exhaustive investigation of the problem was that of Konig & Grossfeld, who in 1913 set out definitely to clear up the discrepancy. As perhaps might have been expected, they found that they could isolate purine bases after hydrolysis from all the fish eggs they studied, but only in small quantity; their results are shown in Table 36. The question of nuclein synthesis by the developing embryo will be discussed in relation to these findings in Section I0'3. Table 36. Investigations of Konig & Grossfeld. Total purine bases isolated Mgm.% dry weight Herring Carp Cod ... Pike ... Sturgeon 0-408 I -060 2-440 0-014 0-230 But the exact relationship between the nuclein and the vitelHn remains exceedingly obscure. It is possible that in one and the same egg there may be more than one modification of vitelHn, apart altogether from the insoluble form suggested by Steudel & Osato. All the knowledge that we possess at the present time 338 THE UNFERTILISED EGG AS A [PT. Ill on this point is of an unsatisfactory histological nature, and any discussion of it must inevitably include an unprofitable proportion of guesswork. Thus, Jorgensen differentiated histologically between two substances which seemed to be present in the unripe egg of Patella vulgata, ergastoplasm No. i and ergastoplasm No. 2, one at least of which was responsible for the formation of the vitelline globules. Faure-Fremiet & Garrault identified ergastoplasm No. i with the mitochondria, and ergastoplasm No. 2 with the fatty constituents of the yolk. But if two forms of vitellin existed, one in loose combination with a nuclein and the other free, the staining reactions of histological elements mainly constituted by one or other of these Table 37. Phos Millon Trypto Glucos Investigator Protein Iron phorus test phane Sulphur amine and date Ichthulin None Much Positive Negative Present None McCrudden (1921) Albumen Traces Present ,, ,j Much ,, Ichthulin None A little Negative )> None >j Albumen ,, Much ,, Present Ichthulin — Present — — None Levene (1901) ,, Present J, — — Present Walther (1891) ^^ — — — — None Hammarsten (1905) " — Present — — — Valenciennes & Fremy (1854) ,, — jj — — >j — Gobley (1850) substances would very likely differ, and it is possible that an explanation on these lines may in the future correlate the chemistry with the histology of the yolk. The vitellin question has been in a measure reviewed by McCrudden, whose table (given in Table 37) illustrates the difficulty of summing up the findings of investigators at all succinctly. The amino-acid analyses (Table 34) are rather more interesting. We have data for the vitellins of the herring, the trout, the cod, and the sturgeon among fishes, the frog among amphibia, the grasssnake among reptiles, and Hemifusus tuba, a gastropod. To this may be added amino-acid analyses of the mixed egg-proteins of the seaurchin egg and the eggs of the brook-trout and the giant salamander, as well as the albumens of cod and sturgeon and the mucoprotein of Hemifusus. If the fish ichthulin analyses of Iguchi or Hugounenq be compared with those of Table 1 1 for the vitellin of the hen, no very marked differences can be observed, although the predominancy SECT, i] PHYSICO-CHEMICAL SYSTEM 339 of arginine and lysine over histidine, which is a constant feature of the ichthulins, reaches greater values in the latter than in the case of bird vitellin (see Table 38). Again, bird vitellin always shows a notable proportion of proline and leucine, and this is also the case with the vitellins of the lower animals (e.g. 10 per cent, of leucine in gastropod vitellin, 19 per cent, in snake vitellin and 9 per cent, in herring ichthulin), though the amount of proline is usually not so great. The only instance of a real divergence between bird and other vitellins would appear to be the glutamic acid content, which is always high in the former, although this amino-acid is absent from the latter. Table 38. Hexone bases of yolk-proteins. In gm. % original In % total nitrogen protein Investigator Species Protein Hist. Arg. Lysine Hist. Arg. Lysine and date Herring Ichthulin 2-45 I4"50 10-07 I'^S 6-33 7-40 Steudel & Takahashi (1923) Egg-menibrane 3-99 14-41 7-51 2-09 6-35 5-55 Steudel & Osato (1923) protein Hen Vitellin — — — i-go 7-46 4-81 Osborne & Jones ( 1 909) Herring Ichthulin 0-40 2-70 2-00 — — — Hugounenq (1904) (clupeovin) Sturgeon Ichthulin 0-47 0-97 o-oi — — — Konig & Grossfeld (1913) Cod Ichthulin 0-55 0-54 0-02 — — — ,, ,, Trout Ichthulin 0-54 0-41 o-oi — — — ,, ,, Gastropod Ichthulin None 3-73 0-86 — — — Komori (1926) Frog Vitellin 1-14 1-06 0-29 — — — Galimard (1904) (ranovin) Snake Vitellin 0-30 0-32 1-45 — — — ,, If now Table 39 is considered, it will be seen that variations are present in the general analysis of these proteins, but that they tend to cancel each other out among the groups. Thus the mono-aminoacid/di-amino-acid ratio is very constant indeed in different ichthulins, although Rothera himself considered that he was dealing with two entirely different proteins, the vitellin of the Torpedo egg and that of the sturgeon. It is unfortunate that Komori's examination of gastropod vitellin was confined to the estimation of the amino-acids by isolation, and did not include a van Slyke determination of the relative amounts of mono-amino and di-amino acids. In contradistinction to the ichthulins, the mixed egg-proteins studied by Russo and Gortner show more variation, though the former's values for two sea-urchin ^gg proteins agree well with the usual vitellin figure. Masing, however, was not able to find any phosphoprotein phos 340 THE UNFERTILISED EGG AS A [PT. Ill phorus in sea-urchin eggs, and Needham & Needham found only very little. It is interesting to note that the ratio is subject to large fluctuations among the keratins of the egg-cases. As for the albumens which Konig & Grossfeld isolated from the eggs of the sturgeon and the cod, they seem to approach in their composition, in so far as data for the hexone bases permit one to form a conclusion, the ovoalbumen in the hen's egg. The 8 per cent, of tyrosine obtained from the sturgeon ovoalbumen is, however, remarkable. The mucoprotein which Komori found around the eggs of the gastropod Hemifusus tuba, and which he partially analysed, is not sufficiently well characterised to be compared except roughly with the mucoprotein of the amphibian egg-jelly. Table 39. In % total nitrogen Species Torpedo {Torpedo marmorata) Sturgeon Dogfish (Scyllium stellare) „ (Pristiurus melanostoma) ,, (Scy Ilium caniculd) Hen Herring ... Sea-urchin Brook-trout Giant salamander Hen Protein Ichthulin Ovokeratin Ichthulin Egg-membrane protein Mixed eggproteins (total) Mixed egg-proteins (coag. only) Mixed eggproteins Vitellin (for i comparison) 15-67 1609 i5-o8 14-33 14-23 1 6-43 14-09 ipz 8-49 849 1-26 6o-20 9-51 5-09 5-13 449 660 i-8i 0-99 0-56 0-14 0-24 0-21 63-60 79-66 66-45 64-19 73-70 61-77 25-10 27-65 15-78 28-78 30-75 2050 27-02 2-40 2-30 5-04 2-31 209 3-55 2-29 Investigator and date Rothera (1904) Buchtala (1908) Steudel & Takahashi (1923) Steudel & Osato (1923) Russo (1926) 7-33 2-05 — 62-11 25-91 2-40 — 284 45-70 17-30 2-64 — — — 62-20 29-80 209 ,, 1-82 — — 61-55 28-25 2-18 Gortner (1913) 2-25 — 1-63 S-55 53-73 29-35 1-83 67-10 25-10 2-67 Plimmer (1908) The general distribution of nitrogenous substances in the eggs of the lower animals is shown in Tables 40 and 41. Pigorini's investigation of the silkworm egg is suggestive, but his data about the different protein fractions are insufficient to enable us to form any judgment on their relation to those so well known in the bird's egg. The very large amount of mucoprotein in the silkworm ovum is certainly remarkable. In Table 41 are placed the few data which we have on the relative amounts of protein and non-protein nitrogen in different eggs, and the way the protein is divided between keratin, albumen, and ichthulin or vitellin. Clearly enough there is great variation, and a rough dichotomy into two groups, one in which the SECT. I] PHYSICO-CHEMICAL SYSTEM 341 non-protein nitrogen accounts for from 14 to 35 per cent, of the total nitrogen, and one in which it only accounts for less than 10 per cent, of the total nitrogen. It is evident from the work of Konig & Grossfeld that all the fishes examined belong to the first of these categories, although within the group there are wide divergences, such as the minute amount of albumen apparently present in the trout's egg and the low non-protein nitrogen of the herring's egg. Good agreement is to be noted between the results of Levene and Konig & Grossfeld, who all worked on the cod; and, although nothing concerning the non-protein nitrogen can be gathered from the figures of Kensington and Hugounenq, their results do show general agreement as regards the partition of nitrogen among the proteins. The only reptile on whose eggs work has been done which could be incorporated in the table is the grass-snake, and there, although no non-protein nitrogen figures are available, it is interesting to note the very high proportion of keratin. Table 40. Silkworm (Bomfryx: mon). (Pigorini, 1Q23.) In % of total protein A Protein sol. in water but not Protein sol. in Protein sol. in Protein sol. in coagulable by heat, and distilled water 10 °„ salt sol. dilute alkalies yielding glucosamine on (albumen) (vitellin) (nucleoprotein) hydrolysis (ovomucoid) 29-20 8-57 11-45 5090 The second principal group, consisting of those eggs which have a relatively much lower percentage of non-protein nitrogen, contains two members, the hen and the silkworm. The former may be said with a high degree of probability to be characteristic of all nidifugous birds, and perhaps of nidicolous ones also, but whether the latter is at all representative of the centrolecithal insect eggs may be considered doubtful. The sole insect egg which has been investigated chemically, so far, is that of the silkworm, and until more evidence is available the hen and the silkworm will have to be placed together in this second group without comment. It is significant that, in the hen's case, the percentage of albumen is greater than in any other, a fact obviously referable to the large amount of egg-white present in that egg. Finally, it is of interest that the sea-urchin's egg seems to have a protein/non-protein nitrogen ratio very like that of the fishes, but situated on the low protein edge of their limits of variation. 342 THE UNFERTILISED EGG AS A [pt. iii Table 41. Distribution of Gm. % wet weight Species Water Protein (Nx 6-25) Protein (by diff.) Protein of eggmembrane Albumen Ichthulin Nitrogen (direct) Free bases and amino acids Fat Ash Carp ... 66-15 2770 2997 363 16 43 4-432 9-91 2-48 1-40 Pike 6353 28-13 33-01 375 2-38 17-29 4-500 9-59 1-40 2-06 Trout 6385 27-81 30-81 1-76 0-15 24-33 4-450 457 3-71 1-63 Herring 69-22 26-32 2521 3-20 4-83 13-68 4-212 3-50 4-19 1-38 Cod 72-10 23-02 24-44 2-57 2-70 11-47 3-683 7-70 1-33 2-13 Salmon — — — — — — — — — Herring 65-00 28-ss — 0-79 28-70 — 3-62 Sea-urchin (Strongylo centrotus lividus) Silkworm 80-50 6-47 35-00 10-20 3-00 3-90 2-25 66-24 22-00 — — — 3-67 0-875 — Grass-snake (Tropido notus natrix) Cod 58-94 94-67 19-24 — 12-71 0-72 5-81 — Fresh- water gar 5390 26-20 — — — — 0-138 Hen (average results) whole egg Turtle (Thalassoclielys corticata) yolk — — 11-81 0-45 6-13 5-23 2-91 0-500 0-033 —

With so % mucoprotein and Within the non-protein nitrogen fraction itself there are some fragmentary data for the distribution, as may be seen from Table 42 . Unidentified compounds usually account for from 20 to 35 per cent, of the total non-protein nitrogen, and free amino-acids for approximately half of it. Among those identified by Steudel & Osato were histidine, arginine, lysine and cystine. The ammonia may vary from 4 to 25 per cent., and the purine bases from 15 to 40 per cent. As far as can be seen at present, the hen's egg seems to possess the greater part of its non-protein nitrogen in the basic fraction. The most interesting point brought out by the table is probably the significant quantity of urea shown to be present by the analyses of Steudel & Osato, amounting to no less than half of the total nonprotein nitrogen, and it is possible that a good deal of the unidentified nitrogen of Konig & Grossfeld might be accounted for in this way. The presence of nitrogenous excretory products in the undeveloped egg, though at first sight paradoxical, is nevertheless undoubtedly a fact in the case of some aquatic organisms. The hen's egg contains hardly a trace of urea at the beginning of development but that of a selachian fish contains a good deal (see Section 9- 1 1 ) . SECT, i] PHYSICO-CHEMICAL SYSTEM 343 the nitrogen in eggs. Gm. % dry weight (ash free) ^ A . ^ % of the total nitrogen Egg , — ^ ■ — \ Pro- mem- Free Free tain Pro- brane Nitro- bases and Protein Ker- Albu- Ichthu- amino- (N x tein (by pro- Albu- Ichthu- gen amino- Investigate total atin men Hn acids 625) diflF.) tein men Hn (direct) acids Fat Ash and date 66-9 121 548 331 85-37 92-37 1119 50-64 13-65 30-54 7-64 — Konig & Gr (1913) 70-96 11-35 721 52-4 2904 82-75 95-93 10-90 6-92 50-25 13-08 27-87 407 — „ 85-20 15-7 0-49 790 1485 80-56 89-25 5-10 0-44 70-48 12-89 13-24 IO-7S — 86-15 12-70 9-15 543 13-9 89-52 85-75 10-88 16-43 46-53 14-33 "-Qi 14-25 — 68-41 10-50 i-oi 469 315 8933 94-84 9-97 10-48 44-51 14-29 29-88 5-55 — „ (loo-o?) — 11-70 880 — 87-80 — — 10-30 77-5 — — 4-50 7-50 Kensington ( (loo-o?) 2-7 973 — __ — — — — — — — Hugounenq 62-0 — — — 379 — _ — — — — — — — Russo (1926) 94-0 — 27-4 8-05 6-04* — — — — — — — — — Monzini(i9: Pigorini (ic 96-0 — — — 3-98 65-25 — — — — — — — — Russo (1922) (loo-o?) 66-0 3-74 30-2 — — — — — — — — — — Galimard (19 66-0 — — — 33-0 68-09 — — — — — — — — Levene (i89( _ ___ — _________ Nelson & Gi (1921) 96-4 406 49-8 42-5 366 — — — — — — — — — — Q8-5 — — — 1-5 — — — — — — — — — Tomita (1921 chorionin in addition. As is well known, these fishes have a special relation to this substance. In 1858 Stadeler & Frerichs isolated "kolossale Quantitaten von Harnstoff" from the organs of plagiostomes, obtaining a solid mass of urea nitrate when they added nitric acid to their final concentrates. One liver of an adult Scy Ilium canicula gave them 2 oz. of urea, and similar high figures were reported for Acanthias vulgaris. Teleostean fishes, however, and the cyclostome, Petromyzon planeri, yielded practically no urea, at any rate not more than would be present in mammalian tissues. Stadeler confirmed the selachian results on Raia batis and clavata and on Torpedo marmorata and ocellata. In 1 86 1 Schulze repeated and confirmed Stadeler's work on Torpedo, and in 1888 Krukenberg published an extensive work on the subject, in which he related his unsuccessful attempts to demonstrate urea in the bodies of teleosts {Lophius piscatorius. Conger vulgaris, Acipenser sturio), a cyclostome [Petromyzon fluviatilis and Ammocoetes) and a cephalochordate (Amphioxus lanceolatus) , although he found large amounts of it in the bodies of elasmobranch fishes [Scyllium stellare, Mustelus vulgaris and laevis, Acanthias vulgaris, Squatina angelus, Torpedo marmorata, Myliobatis aquila) and in the holocephalic Chimaera 344 THE UNFERTILISED EGG AS A [pt. iii monstrosa. Particularly interesting were his experiments with eggs — he isolated considerable amounts of urea from a 5 cm. embryo of Mustelus laevis, and from the yolk of Scyllium stellare and Myliobatis aquila eggs, but he could find none in the surrounding jelly or "white ". An Ggg ofPristis antiquorum yielded 3920 mgm. per cent, (wet weight) and a Torpedo ocellata egg 1 740 mgm. per cent. An Acanthias vulgaris embryo 1 7 cm. long had 3360 mgm. per cent, in its muscles, 1800 mgm. per cent, in its liver, and 2640 mgm. per cent, in its unused yolk. Other work on urea in selachians was done by Grehant and by Rabuteau & Papillon. Table 42. Distribution of non-protein nitrogen in eggs. % of total non-protein N (including purine N) g-| z § iz g 2 -g I I |Z w.« „ „ ^ 2^ c S "G 2 o op Investigator Species HSo? cq <fe^ P U D U hUa and date Herring — 198 — 44-3 359 — — — — — Konig & Grossfeld (1913) Carp ... ... — 39-8 — 36-1 24-1 — — — — — >> >, Sturgeon ... — 25-2 13-6 55-4 189 — — — — — ,, ,, Herring ... ... 2060 244 67 21-6 — 519 None 18-3 — — Steudel & Osato (1923); Steudel & Takahashi (1923) Herring 1443 16-91 23-42 41-65 1802 — — — — — Yoshimura (1913) Silkworm ... 440 — 4-44 54-30 34-60 — - — — 610 6-7 Russo (1922) Hen (aver, figures) — 88-80 4-22 7-04 — None None Trace — — — Fresh- water gar 299 92-00 — 4-02 — — — 4-0 — — Nelson & Greene (1921) (not ripe) More light, however, was thrown on the reasons for this richness in urea when in 1897 Bottazzi working on the osmotic pressure offish blood, found that the elasmobranchs differed fundamentally from teleosts in being isotonic with sea water. Serum A Selachians Torpedo marmorata —2-26° Trygon violacea —2-44° Teleosteans Charax pimtazzo —1-04° Serranus gigas — i -03° Bottazzi observed that the selachian osmotic pressure would correspond to some 3-9 per cent, sodium chloride but laid no emphasis on the fact that selachian blood did not contain anything like so much ash. It was left for Rodier to show that the difference was made up almost wholly by urea. Duval has since found that the salts alone would only give an osmotic pressure of A — i-o6°. "High bloodurea", as Smith says, "is a phyletic character of the orders Selachii SECT, i] PHYSICO-CHEMICAL SYSTEM 345 and Batoidei", and its osmotic function was well shown by the reciprocal relation between salts and urea which Smith found to hold in selachian tissues and fluids. Blood-urea mgm. % Smith (1929) Selachians Dogfish {Mustelus canis) 880 Denis (1913) Selachians Dogfish {Mustelus canis) 800 Sandshark {Carcharias littoralis)... 1000 Skate {Raia erinacea) 868 99 Teleosteans Mackerel {Scomber scombrus) 86 Goosefish {Lophius piscatorius) ... 40 Flounder {Paralichthys dentalus) 46 In view of all these facts it is not surprising that Needham & Needham in 1928 found about 5 mgm. of urea nitrogen present in the Scyllium canicula egg at the beginning of development ; and 888 mgm. per cent, of urea in the undeveloped Acanthias vulgaris egg. Gori, again, found 7 10 mgm. in undeveloped Torpedo eggs. But since urea accumulation is closely confined to elasmobranchs it is unlikely that the results of Steudel & Takahashi and of Konig & Grossfeld can be interpreted as being due to urea. The presence of urea has also been reported in the undeveloped eggs of "ants and flies" (in small quantities) by Fosse. Further details would be desirable here. There is reason to believe that nitrogenous substances other than those already mentioned are present in certain eggs. Thus Yoshimura and Poller & Linneweh isolated trimethylamine, tetramethylenediamine and choline from fresh herring eggs, and there is a certain probability that fish eggs also contain betaine. As the characteristic smell of fish is due to these amines and related substances, this is not very surprising. Brieger is said to have found neuridine in fish eggs, and Schii eking isolated spermine from echinoderm eggs in 1903. Taurine and glycine were found in echinoderm eggs by Kossel & Edlbacher. Of the manner of formation of ichthulin in the maturation of the ovum we know absolutely nothing. Paton & Newbigin concluded from a very few analyses that the phosphorus was brought to the ovaries from the muscle of the salmon as inorganic phosphorus, but, in view of what is now known about the organic phosphorus compounds of blood, this appears rather unlikely. 346 THE UNFERTILISED EGG AS A [pt. in I '14. Fats, Lipoids and Sterols Studies on the fatty substances of the undeveloped eggs of different animals have resulted in much interesting information. There has been, of course, a great body of histological work, and the yolks of all kinds of eggs have been repeatedly subjected to microscopic examination (for example, Kaneko's study on the silkworm); but, in spite of many attempts, I have not succeeded in finding more than a few hints in this literature which are of value to the chemical worker. This subject has been dealt with in a general way by Ransom and by Dubuisson, to whose papers those interested in the histological aspects of yolk must be referred. Of the way in which the fat and the protein are intermingled in the yolk we know practically nothing, and it would be most desirable to investigate the yolk with the methods which modern colloidal chemistry has developed. But that the association between fat and protein indicated by the histological evidence is not very close is shown by the interesting centrifugation experiments of McClendon on the amphibian egg. If the egg of the frog is centrifuged for five minutes under the right conditions, it separates into three perfectly distinct layers, the upper one being oily and yellow, the middle one translucent, colourless and protoplasmic, and the lowest one black, containing practically all the yolk. By using a considerable number of eggs, McClendon was enabled to obtain suflticient material for the chemical analysis of each layer. The figures he obtained are shown in Table 43. It is evident from a slight inspection of his results that the upper layer is composed mainly of neutral fats and a little lecithin, and the middle layer of water, salts and protein, with no fats or lipoids. The lowest and much the largest layer is made up of the vitellin (ranovin or batrachiolin) together with the major part of the lecithin. It is interesting that the association between the phosphoprotein and the lipoid was the only one that centrifuging could not break, for, as we have already seen, the observation of a loose lecitho-vitellin combination in the hen's egg is very old. McClendon found that mitotic figures were all present in the middle layer, and that this centrifuging produced a variety of monstrous embryos. He was led to regard the protoplasm of the egg as constant in composition throughout, but "anisotropic as regards its axes, in other words crystalline in structure". SECT. l] PHYSICO-CHEMICAL SYSTEM 347 McClendon extended his observations to the egg of the sea-urchin, Arbacia punctulata. Separated by centrifugal force, this egg divided itself into four layers, as Lyon had already described, {a) a layer of yolk bodies and red pigment granules extending from the centrifugal end about half-way to the equator, {b) a layer of similar yolk bodies but without the pigment granules, {c) a translucent fluid layer extending almost to the centripetal pole and containing the nucleus, and finally {d) a very opaque layer or cap of minute volume, sitting on the centripetal pole. When the crushed eggs were centrifuged, the material separated into two layers, {a) and {b) being indistinguishable, centrifugal and containing the egg-membranes, and [c) centripetal, {d) not being perceptible. McClendon analysed the layers in the same manner as those of the frog's egg — the figures are given in Table 43. Table 43. McClendon' s figures (1909). In the /o in the layers % , of I" 1 "a lay i ers dry weight phosphorus A ll la Is / t! 2 c t< —1 1) u

-> tJ a & 2 1 Layers of centri o-o6 6 Upper centripetal 50 50 8o-o 4-0 8 8 Trace 0-34 1-4 0-41 — fuged egg of (fatty or oily) Rana pipiens 016 16 Middle (protoplasmic) 82 18 7-5 II-5 60 21 Trace 0-05 i-o 0-37 ~ ~ 0-78 78 Lower centrifugal (yolky) 48 52 24-0 60 10 60 Trace 270 1-2 1-33 ~ ~ Layers of centri — 32-5 Centripetal (proto 88 12 49-0 — 20 3I-0 2 36 16-66 3-24 13-45 1-24 fuged egg of ^r plasmic) bacia punctulata — 67-5 Centrifugal (yolky) 79 21 38-2 — 10 51-8 2 74 12-84 1-6 10-6 2-02 A short consideration of them shows that centrifugal force is not nearly so successful in separating the egg of the sea-urchin into chemically unlike layers as it is in the case of the frog. This fits in perhaps with the long-established fact that centrifugal force interferes far less with normal development in the sea-urchin's egg than it does in the frog's egg (Morgan and Lyon). It was very noticeable that, whereas the frog's egg separated out into layers of markedly different water-content, this did not take place in the sea-urchin's egg. In the case of the centrifuged frog's egg, again, there were big differences between the phosphorus contents of the different layers, but in that of the sea-urchin's egg this only applied to the residues which were mainly protein. McClendon surmised that the inclusion 348 THE UNFERTILISED EGG AS A [pt. iii of the membrane proteins in the centrifugal layer caused this effect. It is of course a fact of the first importance that normal development can follow centrifugation and this will receive attention later (see Section 3 and the Epilegomena) . I shall only mention here as one of the best instances of this phenomenon, the work of Schaxel on the axolotl egg. Here centrifugation caused atypical discoidal cleavage which nevertheless resulted in a normally proportioned embryo. Thus normal conclusions can follow abnormal distribution of the so-called "organ-forming substances". For further details of these experiments, see Morgan and Bertalanffy. The early work of Gobley on the fat of the hen's and the carp's egg has already been described. He isolated glycerophosphoric acid from the latter, and pursued further his investigation of lecithin, concerning which it is of interest to note that Sacc contested his claim to have found organic alcohol-soluble phosphorus. Sacc believed that the fats contained dissolved in them a quantity of inorganic phosphorus. Gobley, however, was easily able to disprove this view and to show the identity of carp's egg lecithin with brain lecithin. Data which have accumulated since Gobley's time on the fatty substances of the eggs of the lower animals are collected in Table 44, and may be compared with those in Table 22. One of the most striking differences between the hen's egg and other eggs is the relatively low iodine value of the fatty acids of the former, both free and combined in lipoids. The neutral fat of the hen's egg has an iodine value varying roughly between 60 and 90, but for fish eggs the figures vary from 90 to 150, and the same rule holds generally of the lipoid fatty acids, for they average 60 in hen and 100 in fish eggs. The saponification numbers, on the other hand, are much the same throughout the two tables (from 170 to 200). The conclusion might therefore be drawn that egg fats differ rather more as to the number of unsaturated linkages in their acids, than as to the length of their chains. Nevertheless, there are remarkable exceptions to these generalities, the fatty acids of the echinoderm eggs, for example, having enormous saponification and high Dyer numbers, and therefore presumably only very short chains of carbon atoms. Arbacia is more remarkable in this than Asterias. Yet, though they are exceptional in that respect, they have iodine numbers very like those of fish-egg fats. Another point of interest is that the cholesterol/fatty SECT. I] PHYSICO-CHEMICAL SYSTEM 349 acid ratio, as shown by expressing the cholesterol in percentage of the fat present, is rather constant, never going below 4 and never rising above 12. This may have some connection with the physical Table 44. Data for fat fraction of eggs. 1 "o-tio a <S 1 3 ■3 a ^1 l^^l J3 0? £ C a ?f H ^ 5 ^ <4-r ."2 ."S "S - c 3 ■$t Qj cj.S .2 ■3 c 0? '0 '0 •s| &3 .-a (u u — 2 J5 4^ 11 'S 2 3 u p Investigator Species ^ ^ m c Q^Z c u^ Ji fc fe cfiJ: D^ and date Amphibia Frog {Rana tempor 123-0 — — 5-97 — — — — — Faure-Fremiet & Dragoiu aria) (1923) Fishes Sturgeon ... 107-6 191-4 — 4-35 12-92 — — — — Konig & Grossfeld (191 3) Trout 128-3 181-8 — 6-52 41-10 o-ig 0-19 — — „ ,, Cod 148-4 1 76- 1 — 12-05 35-19 0-21 0-43 — — ,, jj Herring ... 123-1 230-6 — 6-94 43-61 — — — ,, J, Carp 78-9 186-9 — 10-98 59-19 — — — — 55 55 Pike — — — — — 0-27 0-22 — — 55 3J Trout {Salmo fario) 108-6 219-8 — 6-23 37-50 — — — 1-7 Faure-Fremiet & Gar (132-9 189-8 Fatty acids of the ph osphatide fraction' rault (1922) Carp {Cyprinus carpio) — 140-0 ■ — 5-99 60-00 — — — 3-36 55 55 (64-4 — Fatty acids of the phosphatide fraction' Dogfish (55-88 — Fatty acids of the phosphatide fraction^ Ponce (1924) Shark (Lepidorhinus — — — — — — — 17-3 — Tsujimoto (1920) kinbei) ECHINODERMS Sea-urchin {Arbacia 147-0 606-0 4-001 — — — — — — Page (1927) punctulata) Sea-urchin [Echinus 145-0 195-0 — — 29-40 — — — — Moore, Whitley & Adams esculentus) (78-8 225-0 Fatty acids of the phosphatide fraction] (1913) Sea-urchin {Arbacia — — — — 50-00 — — — — Matthews (191 3) punctulata) Sea-urchin [Paracen 150-0 — — — — — — — — Ephrussi & Rapkine trotus lividus) (1928) Starfish (Asteriasgla 112-5 318-8 3-778 — — — — — — Page (1927) cialis) POLYCHAETE Polychaete worm — — — 8-85 — — — — IO-6 Faure-Fremiet (1921) {Sabellaria alveolata) Nematode Roundworm {Ascaris — — — 5-0 — — — — 80-0 Faure-Fremiet (191 3) megalocephala) State of the egg-cell, and will be referred to again (see Section 12-5). The lipoids, expressed as lecithin in per cent, of the fat present, show greater variations, but it is not possible to say at present what the significance of these may be. 350 THE UNFERTILISED EGG AS A [pt. iii The mention of squalene in Table 44 indicates the existence of an egg-constituent, our knowledge of which is of very recent origin. In 1 906 Tsujimoto isolated from the liver oils of elasmobranch fishes a saturated hydrocarbon of approximate formula C30H20, and in 191 6 published a further study of it. Its properties and constants are given in Table 22. In 1920 he reported that he had been able to isolate it from the egg-yolks of two elasmobranchs, Chlamydoselachus anguineus and Lepidorhinus kinbei, where it made up no less than 13 per cent, of the egg (wet weight) and, in another case, 1 7 per cent, at least of the total fat fraction. There the matter rested until 1926, when Heilbron, Kamm & Owens, taking up the question of its presence in eggs once more, isolated it from the undeveloped yolks of Etmopterus spinax, Lepidorhinus squamosus and Scymnorhinus lichia. In the fully developed eggs of the first-named of these three, practically none was present, indicating that it must either have been combusted or absorbed during development. Further researches on the embryological significance of this compound are greatly required. It is possible that some hydrocarbon of this sort may explain certain obscure points in the chemistry of the egg, for instance, the oil extracted by Dubois from the locust's egg {Acridium peregrinum). It contained 1-92 per cent, phosphorus, and was present to the extent of 4*5 per cent, of the wet weight of the egg, no small proportion. Kedzie studied a similar oil which he obtained from the egg of the American locust. A question which is perhaps related to the general problem of the egg-oils is that of the oil-globules of the yolks of some of the teleostean fishes. In 1885 Agassiz & Whitman divided all pelagic eggs into those which had the oil-globule and those which had not. But it was soon found that this method of classification was valueless, for the appearance of the globule is rather erratic; thus, although Lota vulgaris (van 'S>2ivs\he\ie) , Brosmius (anon.) and Motella mustela (Brook) were all found to have it, the common pike's egg does not have it (Truman). Ryder first suggested that the oil-globule might have a relation to buoyancy, but Prince, reviewing the whole subject a little later, pointed out that this could hardly be so, for the salmonoid fishes all have them, and yet their eggs never float. Moreover, out of 22 teleost eggs with no globule, 17 are pelagic, while out of 24 teleost eggs which have globules, only 15 are pelagic. Ryder replied to this by partially withdrawing his theory, and Mcintosh SECT, i] PHYSICO-CHEMICAL SYSTEM 351 simultaneously showed that the eggs of the catfish, which are undoubtedly bottom ova, have large oil-globules. Another theory was put forward by van Bambeke, who believed that the oil-globule was a special form of yolk, and of a purely nutritional significance. Prince criticised this view on the ground that the oil persists in the yolk after the liberation of the embryo from the egg-membrane, and travels beneath it as it swims about. This would not, however, negative the possibility that the oil was used for larval rather than embryonic nourishment. Van Bambeke' s claim that a protoplasmic thread passes from the oil-globule to the germinal disc was almost completely disproved by van Beneden. His and Miescher, examining the oil histochemically, found that it only stained very slowly with osmic acid, and therefore differed profoundly from the yolk, and, although it was soluble in ether, it contained no more than a trace of phosphorus. It is remarkable that the oil has never been subjected to a proper chemical examination, especially in view of the extensive zoological literature on it. What we know of its properties faintly hints, perhaps, that it may be a hydrocarbon like squalene, and the whole question, indeed, holds out great possibilities for physiological as well as chemical work. The oil must readily dissolve lipochromes, for the pink pigment of the salmonoids is found in it. Prince's own theory was that the globule was a constituent of ancestral significance, a vestige from the time when, as Balfour showed, the teleostean yolk was very much larger than it is now. The nutrition view is probably the best. The lipoids and sterols of the eggs of the lower animals are very little known, and their further study is much to be desired. Page in 1923 described a sterol — asteriasterol — which he isolated from the eggs of Asterias forbesii and which turned out to be closely related to, though not identical with, ordinary- cholesterol; the eggs of Arenicola cristata, on the contrary, yielded a sterol absolutely identical with the well-known substance as it occurs in mammals. Ten years previously, in a less accurate study, Matthews had failed to find any cholesterol at all in the eggs of Asterias forbesii, though he had been able to isolate some from those of Arbacia punctata. From the former he got a jecorin-like substance, containing 10 per cent, of glucosamine, which was probably a mixture of kephalin, cerebrosides, "protagon" and various carbohydrates. Page's later study of the fats and lipoids of the echinoderm egg led to the conclusion that (qualitatively) there was more kephalin in the eggs of Arbacia than 352 THE UNFERTILISED EGG AS A [pt. iii in those of Asterias, and more lecithin in the eggs of Asterias than in those of Arbacia. Asterias contains large amounts of soaps, and its oil is present in much greater abundance than the oil of Arbacia; moreover, it contains more sulphur compounds (sulphatides?) decomposable with potash than does the Arbacia egg. Page, Chambers & Clowes made a study of the effects of various cytolytic agents on the eggs of Asterias separated by microdissection into their cortical and endoplasmic components. They used for this purpose hypotonic sea water, digitonin and saponin, and found that digitonin caused slow cytolysis of the cortical and rapid cytolysis of the interior protoplasm when the two were isolated, whereas hypotonic sea water caused slow cytolysis of the interior and rapid cytolysis of the cortical protoplasm. If these results do not actually demonstrate that the greater part of the asteriasterol is localised in the outer and fertilisable parts of the egg, they at any rate suggest a new method of investigation which may help to solve many similar questions in the future. Runnstrom has studied the lipoids of the echinoderm Qgg in relation to its coloured interference fringes and its membrane properties. Among the sterols existing in eggs must be mentioned a substance which has long been known to occur in the ova of Ascaris, and which has been called "ascarylic acid". Faure-Fremiet identifies it with the droplets or crystals described in the egg oiAscaris by van Beneden. It was isolated simultaneously by Faure-Fremiet from the eggs and by Flury from the whole body of the nematode ; the former worker found that it accounted for 22 per cent, of the dry material. Ascarylic alcohol, ascarylic acid, or, as it would probably be best to call it, ascaristerol, seems to exist in the egg-protoplasm in combination with palmitic, oleic, and perhaps stearic acid in ester form. FaureFremiet & Leroux studied its properties, and proposed the provisional formula of C32Hg404 . Its saponification number was 199, and its m.p. 82°, it did not give the cholesterol colour-reactions, and its molecular weight was close to 511. Flury considered it to be related to oenocarpol. Acaristerol seems to be strictly confined to the eggs, for even the parietal cells of the ovary and uterus do not contain it, as Faure-Fremiet showed by means of histochemical tests. Nor is it present in the testes and spermatozoa. It may at present be classed with the sterols, like asteriasterol. Ascaris eggs also contain o-i6 per cent, dry weight of ordinary cholesterol. SECT. l] PHYSICO-CHEMICAL SYSTEM 353 Table 45. Distribution of phosphorus. In % of the total P f _3 u* Sh u i "o gfln 15 p-^ ^Oh Oh \-i ^ U ^ JJ . w A cx 1^ M)-? £P3 3 3 fl 2 G 3 Investigator Species HDh Hi OI -Si 7 -S fl and date Bird Hen 61-4 9-5 9-5 None 1-6 27-5 29-1 Plimmer & Scott (1909) Amphibian Frog (ovarian) 26-2 4-3 4-3 None 7-6 61-9 69-5 Plimmer & Kaya (1909) Fishes Sturgeon 28-8 i6'9 — 9-9 None 54-3 54*3 Plimmer & Scott (1908) Herring — — — — 63-0 63-0 ■>■) J3 Grey mullet — — — — — 48-0 48-0 5» J> Trout 26-0 — — — — 34-6 — Faure-Fremiet & Garrault (1922) Herring 33-2 — — — — 66-8 — Yoshimura (191 3) Haddock — — 21-22 — — — Milroy (1898) Herring — — — — lO-O 90-0 — Tschernorutzki (191 2) Salmon 37-8 i8-9 — Traces — 43-3 — Paton (1898) Herring — — + + — — Steudel & Takahashi (1923) ECHINODERMS Sea-urchin {Strongylo 43-0 33-1 — — — — 23-8 Robertson & Wasteneys centrotus purpuratus) (1913) Sea-urchin {Arbacia — — — — Much None — Masing (1910) punctulata) Sea-urchin [Arbacia 29-2 47-0 — i-i — — 23-5 McClendon (1909) punctulata) Sand-dollar {Dendras 6-09 46-95 17-55 29-4 32-0 I2-I 44-1 Needham & Needham ( 1 930) ter excentricus) Starfish {Patiria mini 32-3 51-8 32-0 20-0 15-40 Trace 15-4 J> 59 ata) Crustacea Sand-crab (Emerita 28-2 6 1 -40 42-10 19-3 10-82 Trace 10-82 >J J> analoga) Brine-shrimp [Artemia salina) Gephyrea 5-9 56-4 38-3 18-1 37-9 None 37-9 )> >5 Gephyrean worm ( Urechis caupo) Nematode 27-4 56-90 40-60 16-3 15-80 Trace 15-8 5> J) Roundworm {Ascaris) 25-6 — — 20-0 54-4 — 54-4 Faure-Fremiet (191 3) Miescher (1872) reported that in the salmon nearly all the phosphorus is in organic form.

This fraction will include pyrophosphate P. t This fraction will include guanidine phosphoric acids (arginine or creatine phosphate P). + "Present in some quantity." The lipoids of the mammalian egg-cell have recently been the subject of some work which is interesting, though, Uke all histochemical studies, very difficult to appraise. Following on Russo's claim to have found two different sorts of eggs in rabbits, varying N E I 23 354 THE UNFERTILISED EGG AS A [pt. m in their reactions to staining methods, Fels in 1926 confirmed this difference for the human egg-cell, some specimens of which showed a strong lipoid-reaction (Ciaccio and Smith-Dietrich methods) in the nucleolus while others did not. Fels' illustration is certainly striking. Leupold had already put forward the view that eggs whose nucleoli were rich in lipoids produced females, and the remainder males, but all the evidence, however, is against sex-dimorphism in the mammalian egg (see Parkes' review). These observations, together with those of Pollak on the presence of Reinke's crystals in the egg of Macacus rhesus, and similar work by Limon and" von Ebner (on Cerrus capreolus), are all that we have on the chemical constitution of the mammalian egg-cell. Closely connected with the lipoids of the egg is the distribution of phosphorus compounds in it and Table 45 gives what is known upon this subject. It is interesting to see how the phosphoprotein phosphorus varies, in some eggs being very large in proportion to the total phosphorus, in others being almost insignificant. Masing was wrong in saying that the echinoderm egg has none at all, for Needham & Needham in 1929 observed quite a high percentage in the &gg of the sand-dollar. It is significant in view of what has already been said about the pre-eminence of birds in storing fat in their eggs, that the hen's egg has 20 per cent, more phosphorus in lipoidal form than any other egg investigated. The fishes rank in this respect with the echinoderms and annelids, little diflference being noticeable between alecithic and lecithic eggs. Perhaps this famous distinction involves neutral fat rather than lipoids. It is to be noted from Table 45 that the inorganic phosphorus content of eggs is very variable; in many cases almost none is present, but the haddock's ^gg seems to have no less than 20 per cent, of the total phosphorus in this form. About the same proportion is present in the nematode egg, ii Ascaris can be taken as representative. FaureFremiet was able to identify the calcium phosphate in the egg-cytoplasm with the "hyaline balls" described by van Beneden, using various histochemical reactions (McCallum, Prenant, etc.). Pure calcium phosphate, according to Faure-Fremiet, accounts for 0-4 to 0-6 per cent, of the dry weight of Ascaris eggs, an inconsiderable amount in view of the share it takes in the appearance of the cytoplasm as a whole. SECT, i] PHYSICO-CHEMICAL SYSTEM 355 1-15. Carbohydrates The carbohydrates of the eggs of the lower animals have been less investigated than anything else — a summary of our quantitative knowledge concerning them is shown in Table 46. The presence of glycogen in insect and mollusc eggs was noted by Bernard and in those of arachnids by Balbiani. For the reasons mentioned above, it is difficult to know how trustworthy the figures for carbohydrates are, so bad have the methods been in the past. Faure-Fremiet & du Streel's figure for the glycogen of the frog's tgg must surely be too high, for most of the other workers are agreed on a value of about 2 gm. per cent, wet weight. In the case of animals other than amphibia, the figures are too scattered to permit of any generaUsation: thus, though glycogen was not found in herring's eggs by Steudel & Osato, Gori did note its presence in Torpedo eggs, and the eggs of the reptiles Vipera aspis and Elaphis quadrilineatus, in addition to free carbohydrate. Steudel & Osato pointed out that many histologists such as Goldmann had published results concerning the fish egg which might lead one to suppose that very large amounts of glycogen were there. That this was not found by chemical methods ought to induce, they felt, a more cautious attitude towards histochemical work than was customary; indeed, much of what is called glycogen histochemically can certainly not be glycogen. Greene's carbohydrate figures for the eggs of the king-salmon Oncorhynchus tschawytscha, from Cahfornian rixers, were of special interest, for, throughout the maturation period, the carbohydrate content of the egg remained the same. The duration of the fast did not affect it at all. Quahtative investigations of carbohydrate in eggs have been made by Anderlini on the silkworm egg and by Konopacki, who observed the presence of glycogen microchemically in the perivitelline fluid of the frog's tgg. As has already been mentioned, a carbohydrate group is undoubtedly contained in the mucoprotein of the amphibian egg-jelly, and von Furth's analysis of the egg-cases of the squid Loligo vulgaris showed that their protein also contained a carbohydrate, but whether these play any part in the sugar supply for the developing embryo remains an obscure point. Haensel found an amount of glucose in the frog's tgg which is shown in Table 46, but he also tried the effect of keeping the eggs in solutions of various mono- and di-saccharides, 23-2 356 THE UNFERTILISED EGG AS A [PT. Ill Table 46. Mgm. % wet weight Mgm, % dry weight . ' , , ' ^ 6 ^ 4 ^ % y ^ h y c S-o iJ ^ l-S.^ >■ Investigator Species H-c'fa O h-o^ta O ^nd date Reptiles Turtle ( Thalassochelys corticata) White ... ... ... — Trace — — — — Tomita (1929) Yolk ... ... ... — 100 — — — — ,, Tortoise {Testudo graeca) ... — 140 — — — — Diamare (1910) Amphibia Frog {Rana esc. and fuse.) ... — — 2,520 — — — Kato (1909) ,, ,, ... — — 1,100 — — — Athanasiu (1899) Frog {Rana temp.) ... ... — — — — — 7810 Faure-Fremiet & Dragoiu (1923) Frog {Rana esc. and fuse.) ... — — 2,500 — — — Bleibtreu (1910) Frog {Rana temp.) ... ... — — — — 193 — Gori (1920) ,, ,, ... ... — — 10,140 — — — P'aure-Fremiet & Vivier du Streel (1921) „ ,, ... ... 604 — — 1906 — — Needham (1927) ,, ,, ... ... — — 2,528 — — — Haensel (1908) ,, ,, ... ... — — 1,650 — — — Goldfederova (1925) Fishes Herring — 500 — — — — Steudel & Osato (1923) King-salmon {Oncorhynchus — 96 — — — — Greene (1921) tschawytscha) Trout {Salmo fario) ... ... — — 340 — — — Faure-Fremiet & Garrault (1922) ECHINODERMS Starfish {Asterias glacialis) ... — — 20 — — — Dalcq (1923) Sea-urchin {Echinus esculentus) — — i ,360 — — 8980 Moore, Whitley & Adams (1913) Sea-urchin {Strongyloeentrotus 1360 — — 543° — — Ephrussi & Rapkine lividus) (1928) Insects Silkworm {Bombyx mori) ... — — 1,110 — — — Pigorini (1922) ,, ,, ... — — 1,980 — — — Tichomirov (1882) ), „ ... — — — — — 3080 Vaney & Conte (1911) Bee {Apis mellijica) ... ... — — 2,500 — — — Straus (191 o) Cephalopod Octopus {Sepia officinalis) ... — — None 3620 1000 None Henze (1908) ,, As glucose ... ... 2700 — — — — — ,, ,, As pentose ... ... 2380 — — — — — ,, Polychaete Polychaete worm (^aie/Zana — — 1,270 — — — Faure-Fremiet (192 1) alveolata) Nematode Roundworm {Ascaris megalo- — — — — — 2105 Faure-Fremiet (1913) cephala) 356* Table 47. Ash content of eggs. WllOLB EOU H m I'ARTB Pike (£j« lueiw ) Sturgeon [Acitie Sea-urdiilt [Arb ■jntlurw) Starfuli LUUtia. Sea water (Wo ih Hole) Sea waier {Clia tnga cxpcdilion) Sea- urchin (Sir neylQcentrolus lividtu) Dogfish {^p'""' Spider-crab (A/ , cmicula) ... no vtTTUCOsa) ... Octopiu tStpia Ifianalu) Hen (6'<i//uj Join '»""")■■• MotluK ( Volula Dogfuh {Squaiin otonllUas) Frog {Rana Umporaria) l-rout (&ilmo fontimlii) Salmon {Saimo ndar) ... Wrauc (Labrax lupus) ... Torpedo [Torptdo dctllala) Dogfiah [Snilium canicula) Spidcr-aab {Maia vnrucosi Octopu) (Stpta ojficinalij) a pmtuloia) Gephyrcon worm {Sipuncutus nitdui) . Lugworm [.irtnitola claparti"" Herring {Glupea hariagus) Waicr-sol. Salmon .. Plaice Dogfuh \Squatut acanthiai) (cgg-jclly) Sawfish {Pritlii aniiquorum) Mtlhi:quivalcnLa prcx SO, PO, a — 55 Na Mg Ca V»7 01Q7 413 < .V»4 ■1 325 '17 9-8o 300 10 — I -05 4 107 — 0-99 46-6 - 5 i 1-04 z 1-04 16-29 ^fj^ ■6'65 ■iS-29 3-9B g-6i 11-51 1707 ^bs — 10-79 ^3^ 4<'85

&.Grossreld(i9i3) 2-08 '■«<> Page (1927) a-27 — ■ O4-0 7S-3i 409 ia-37 „ 0-004 5'-7 45-^2 Si-s^ 54*42 57-'i — 4-64 — 34-a 54-03 47 b'^'b'^ 5^'^ Diiunar, ualUnger, vol. 1 ■64 13-44 9"3i I3'5i Goblcy {1850) - — — — Wetzel (1907) — — 33-7 t' z 'U *;? ■54 — afj-g — 3'-9 — '"3 — 559 5 — — 7a — — 7-18 ,8-, 11) b-,!, 4-4 lo/i 10-31 13-93 3-3 11-33 'i-7«  6 — — Silkworm iBombyx n Turtle [Timloiiochew coTticata): Whole egg While Yolk >9'03 30-S8 Btalasccwicz (1926} 6i-oi 63ja RoBb & Correa (1927) McCailum (1926) Greig {1898) Milroy [1898) Perugitt (1879) KrukeabcTS (1888} Uf^UID OP TIIB ' Hen {Callus domaliats) Frog {Rana lembotana) Trout {Salmo/ontinatis) Torpedo {Torptdo ocillata) Spiaer-crab ^oifl wrrwfwfl) ... — Sea-urchin (raracentntus lividus) Octoput [Stpia qfficimlii) Schroder (1909} Karaahima (igag) i?-j Bialasccwiti (i< SECT, i] PHYSICO-CHEMICAL SYSTEM 357 to see if they would grow richer in glycogen. They all did; in fact, he was able to double their glycogen content by this simple means (glucose acted better than sucrose, sucrose than lactose, and lactose than glycerol, though even the latter substance gave an effect) . These curious observations have never been confirmed, and can hardly be said to carry conviction as they stand. Diamare obtained discordant results in his researches on the sugar of various eggs ; thus, he got a rather low value for the free glucose of the egg of Testudo graeca, but none at all, either free or combined, from the eggs of Scyllium catulus or Torpedo marmorata. No explanation can be given for this fact. In connection with carbohydrates, it should be remembered that viper venom, which is in all probability a glucoside, has been shown by Phisalix to be present in active form in the yolks of viper eggs. I -16. Ash We come now to the inorganic substances of eggs. Iron has been shown to be present by microchemical tests in many eggs, such as those of Limnaea, Tubifex, Rana esculenta (where it is massed at the light ventral pole) and Pisidium, by the work of Schneider. Dhere found traces of iron and copper in the eggs of Sepia. Warburg found 0-02 to 0-03 mgm. iron per 100 mg. nitrogen in the unfertilised sea-urchin €:gg', part of it seemed to be in ionic form and part not. According to Wilke-Dorfurt, there are 4-8 mgm. per kilo iodine in oyster egg-shells. Ash analyses of eggs have been made by several workers, whose results, it may be remarked, would have been more easily comparable if they had expressed them in the same way, instead of in nine or ten different ways, omitting in some cases the figures which would enable them to be calculated into a form comparable with each other. Table 47 summarises what is known about the distribution of inorganic substances in eggs. It has entailed a good deal of calculation, for only one of the previous investigators expressed his results in terms of millimols and milliequivalents, and unless this is done it is impossible to gain any idea as to the relative preponderance of cation and anion. The first thing which should be noted is the fact that, when the salts are expressed in per cent, of the total ash, potassium is always there in greater amount than sodium, and nearly always to a greater extent than any other metal. This seems to be quite characteristic of the ovum, though in other systems of SECT. I] PHYSICO-CHEMICAL SYSTEM 357 to see if they would grow richer in glycogen. They all did; in fact, he was able to double their glycogen content by this simple means (glucose acted better than sucrose, sucrose than lactose, and lactose than glycerol, though even the latter substance gave an effect) . These curious observations have never been confirmed, and can hardly be said to carry conviction as they stand. Diamare obtained discordant results in his researches on the sugar of various eggs ; thus, he got a rather low value for the free glucose of the egg of Testudo graeca, but none at all, either free or combined, from the eggs of Scyllium catulus or Torpedo marmorata. No explanation can be given for this fact. In connection with carbohydrates, it should be remembered that viper venom, which is in all probability a glucoside, has been shown by Phisalix to be present in active form in the yolks of viper eggs. i-i6. Ash We come now to the inorganic substances of eggs. Iron has been shown to be present by microchemical tests in many eggs, such as those of Limnaea, Tubifex, Rana esculenta (where it is massed at the Ught ventral pole) and Pisidium, by the work of Schneider. Dhere found traces of iron and copper in the eggs of Sepia. Warburg found 0-02 to 0-03 mgm. iron per 100 mg. nitrogen in the unfertiUsed sea-urchin egg ; part of it seemed to be in ionic form and part not. According to Wilke-Dorfurt, there are 4-8 mgm. per kilo iodine in oyster egg-shells. Ash analyses of eggs have been made by several workers, whose results, it may be remarked, would have been more easily comparable if they had expressed them in the same way, instead of in nine or ten different ways, omitting in some cases the figures which would enable them to be calculated into a form comparable with each other. Table 47 summarises what is known about the distribution of inorganic substances in eggs. It has entailed a good deal of calculation, for only one of the previous investigators expressed his results in terms of millimols and milliequivalents, and unless this is done it is impossible to gain any idea as to the relative preponderance of cation and anion. The first thing which should be noted is the fact that, when the salts are expressed in per cent, of the total ash, potassium is always there in greater amount than sodium, and nearly always to a greater extent than any other metal. This seems to be quite characteristic of the ovum, though in other systems of 358 THE UNFERTILISED EGG AS A [pt. iii the organism other relations are found; thus corpuscles and plasma of some mammalian bloods have converse potassium/sodium ratios, and, as a general rule, potassium preponderates in cells while sodium preponderates in media. Of the anions PO4 usually takes up much the greatest part, but SO4 may in certain cases equal it. In the columns on the right of the table the total anion and total cation are shown, in each case calculated as millimols and as milliequivalents, the former giving an idea of the total number of molecules present, the latter of the total number of valencies. Study of the anion/cation ratio expressed as milliequivalents per cent, wet weight provides an important key to the constitution of the egg, for it shows roughly to what extent anion or cation is held in combination with protein or lipoid, or other organic substances. We have already seen that in the case of the hen's egg, taking both yolk and white into account, the anion/cation ratio is more than unity (Bialascewicz's figures give 2-17), showing that a quantity of sulphur and phosphorus is in organic combination — a conclusion which fits in admirably with all that we know of the hen's egg from other sources. The same relationship is seen in the figures of Konig & Grossfeld for the three fish eggs they investigated, the pike, the cod and the sturgeon. On the other hand, the figures of Page for two echinoderm eggs give ratios much less than unity, demonstrating the organic combination of a good deal of the cation. It may be noticed that the analyses of Dittmar and Page for sea water give ratios in the very close neighbourhood of unity, as would be expected, and indicate at the same time that the ratio cannot be regarded as significant to less than o-og. From what has been said, therefore, it might be concluded that the yolk-laden eggs of the fishes, like that of the hen, have a ratio above unity, while the alecithic echinoderm eggs have ratios much below it. But there are exceptions to this generalisation. The ratio of unity for the carp egg which is given by Gobley's results may perhaps be neglected, owing to the date of the work (1850), and the similar value obtained by Roffo & Correa on a gastropod egg may also be regarded as suspicious because of the enormous amount of sodium chloride that appears in their analysis. But the careful work of Bialascewicz in 1926 does not altogether support the generalisation. His figures for the fish egg are in good agreement with those of Konig and Grossfeld, but his anion/cation ratios for the echinoderms SECT, i] PHYSICO-CHEMICAL SYSTEM 359 do not go below unity, though they approach it much more nearly than do the fishes. Further work is needed to clear up this contradiction. In one case, however, Bialascewicz got a ratio below unity, that of Arenicola claparedii, so that in a general sense his investigations are not opposed to those of Page and Konig & Grossfeld. McCallum's low ratio for the egg of the herring is difficult to explain, but Perugia's analysis of the egg-jelly of ovo viviparous selachians fits in well enough with the majority of the other evidence. Attention might also be drawn to Bialascewicz's high ratio (15) for the eggs of the octopus. Sepia, which would appear to be extraordinarily poor in metallic ions (cf. p. 317, Section 13 and the Epilegomena). Some further light is perhaps thrown on the inorganic composition of eggs by Wetzel's figures for insoluble and soluble ash. He submitted the eggs of various animals to examination, with the following results : % dry weight Species Total ash Insol. ash Sol. ash Sea-urchin {Strongylocentrotus lividus) ... 9-7 2-4 7-2 S^iideT-cvah {Maia squinado) ... ... 4-12 0-27 3-8 Octopus {Sepia officinalis) ... ... ... 2-2 0'59 i'6 Tio^sh. [Scyllium canicula) ... ... ... 5-5 1-15 4-3 In all cases he found more soluble than insoluble salts, i.e., more chlorides than sulphates and phosphates. Table 48. Bialascewicz' s figures. Concen Vol. of tration Vol. of liquid CI in I c.c. Total CI inter c.c. of after Deeree ultra in ultra micellar yolk dilution of filtrate filtrate fluid per Species taken (c.c.) dilution (mg.) (mg-) I c.c. yolk Hen (yolk) 4-8 10 2-o8 I -08 1-080 — >3 4-8 20 4-17 0-472 0-944 0-541 >> 4-8 30 6-25 0-306 0-918 0-569 4-8 40 8-33 0-222 0-888 o-537^ »J 4-8 50 10-40 0-195 0-975 (0-754) It is very interesting, as Bialascewicz points out, that the mineral composition of terrestrial and aquatic animals should be so alike. The preponderance of potassium which is seen in the hen's tg^ does not change as one passes to organisms laying their eggs in an environment containing far more sodium than potassium. Thus, although 36o THE UNFERTILISED EGG AS A [pt. iii the normal sea water has twenty times as much sodium as potassium, fish eggs often have quite twenty times as much potassium as sodium. There would not appear to be in this connection any difference between homoio-osmotic and poikilo-osmotic aquatic animals. It is also obvious from Table 47 that aquatic eggs often have very much less salt in them than the ambient medium, and this would be a special case of the phenomenon found in all marine animals, and termed by Fredericq "Mineral hypotonicity". Bialascewicz arranged the animals he studied in a list of ascending concentration of metalHc ions as follows : Metal gramions Species per litre Octopus {Sepia officinalis) ... o-oi6 Gephyrean worm {Sipunculus nudus) 0-064 Spider-crab {Maia verrucosa) 0-079 Wrasse {Labrax lupus) 0-091 Herring {Clupea harengus) (McCallum) o-ioo Dogfish {Scyllium canicula) ... 0-107 Sea-urchin {Arbacia pustulosa) 0-159 Sea-urchin {Paracentrotus lividus) ... o-i8o which would also be an ascending table of taxonomic groups, were it not for the high metal content of the echinoderm eggs, which exceed even the fishes. There are other points concerning the relative amounts of salts in the eggs which require mention. McCallum, who had for a long time previously been studying the proportion of salts in the ash of animals and parts of animals with reference to the composition of sea water both now and in earlier geological epochs, made an analysis of herring's eggs in 1926. He had previously differentiated between palaeo-chemical salt ratios in bloods, namely, ratios resembling that which pre-Cambrian sea water can be calculated to have possessed, and neo-chemical salt ratios, namely, ratios resembling the sea water of the present day. Thus Limulus polyphemus and Aurelia flavidula, the king-crab and the medusa, which have always been marine animals, now approach the modern sea in the composition of their vascular fluids, but the lobster Homarus americanus, the selachian Acanthias vulgaris, the frog, dog, and man, for instance, all have ratios resembling the composition of the sea water at the appearance of the protovertebrate form. He had also identified the kidney as the organ responsible for maintaining the palaeo-ratios in the salts of the blood. In order to explore the possibility of identifying a palaeo-ratio SECT, i] PHYSICO-CHEMICAL SYSTEM 361 in the contents of the cell itself, he had recourse to eggs, and for those of the herring obtained the following distribution : Ratios on the basis of Na 100 Na K Ca Mg CI 100 216-7 ii'4 18-7 356-8 This stood in marked contrast not only with the vertebrate bloodplasma but also with the Archaean sea water calculated for the time at which life first began to appear in it, thus : Vertebrate blood-plasma (dog) 6-6 2.8 0-7 139-5 Archaean sea water 100 100-250 10 0-05 But after extraction of the dried eggs with water in a Soxhlet apparatus, the determination of the ratio of salts in the soluble part gave results more like the ratio for the Archaean sea water: 100 219-9 5'6 1-6 359-2 McCallum therefore concluded that the soluble part of the ash of the herring's egg exhibits a palaeo-chemical ratio. The bond shown here between the metals and the organic substances is useful in reminding us that even in fish eggs, where the anion/cation ratio is well above unity, some of the metal as well as the acid radicles may be united in organic combination. The relation between the salts in the intermicellar fluid of yolk and those in the dispersed phase itself has been studied by Bialascewicz and by Vladimirov. Bialascewicz worked firstly with the yolks of Torpedo eggs, but also with those of the hen and the trout. He prepared series of mixtures of the yolk with diluents in different concentrations, such as isotonic solutions of lithium sulphate and lithium nitrate, or in some cases distilled water, and then, submitting the mixtures to ultra-filtration, he estimated the ash and its composition in the filtrate and the residue. He first found that the percentage of chlorine bound to the dispersed phase in the ooplasm was practically independent of the degree of dilution, and from this fact he was able to calculate the volume of the intermicellar fluid of the yolk (see Table 48). For the hen's egg this was 0-549, per 362 THE UNFERTILISED EGG AS A [pt. m wet substance, and for the egg of Torpedo ocellata a similar calculation, based on cryoscopic experiments, gave a value of 0-482. On the basis of these figures, he proceeded to study the partition coefficient of each individual ion as between dispersed phase and intermicellar liquid. In Table 49 these partition coefficients are given; they represent the ratio amount of ion in the continuous phase or intermicellar liquid j amount of ion in the dispersed phase. It will be noted from Fig. 17 that as dilution of the original yolk goes on the ratios in some cases change, but in others remain constant. Thus the chlorine of the trout and the hen egg yolk remains constant at 0-5 in the latter and 1-02 in the former case, showing that Table 49. Bialascewicz's figures. a l-H 5 31 i §1 si 34 Is ►3t ■f-s SI 5 .2 ^ ^1 11 •2 ~ K 0-722 I -000 0-890 0-768 I -000 0-870 0-967 I -000 0-945 o-8oo Na 0-942 0-567 0-509 0-331 0051 — 0-080 0728 I- 000 I -000 Ca 0-093 0-391 0-274 0-169 0-321 0-760 0-474 0-696 0-505 I- 000 Mg 0-295 0-460 0-321 0-380 0-157 0-410 0-707 0-631 0-272 0-491 P 0-025 0-244 o-ioo 0-275 — — 0-040 0-318 o-i86 — CI 0-555 0-905 I -000 0-567 0-943 I -000 0-970 I- 000 I -000 0-766 it is very stably combined in the dispersed phase, though in different proportions according to the animal. Thus there is considerably more chlorine in the dispersed than in the continuous phase of the yolk of the avian egg, while in the fish egg there is a very slight excess of chlorine in the continuous phase. In all other instances, however, both as regards the hen and the trout, the excess of ion is in favour of the dispersed phase, the colloidal aggregates of which may therefore be looked upon as reservoirs of ash. Nevertheless, there is a good deal of the sodium combined in the continuous phase, and not a little of the potassium, though here the trout differs from the hen, for the potassium ratio is about 0-9 in the former case and only 0-7 in the latter. All the other ions have lower ratios than these; magnesium, calcium and phosphorus, for instance, are all present to a much greater extent in the dispersed than in the continuous phase. These experiments show also exactly how firmly the ions in the dispersed phase are bound there, and with what ease they may be washed out into the ultra-filtrate. It is apparent from SECT. l] PHYSICO-CHEMICAL SYSTEM 363 S 2 °'^^ £ oT 0-8 it'-r CO) 0-63 « E 0-2 0-lb-. 3sl Bialascewic ^ -sSf. Fig. 17 that the phosphorus, chlorine, and probably sodium in the dispersed phase, are intimately united there, for, however great the dilution of it, they do not increase in the ultra-filtrate. Magnesium and calcium, on the other hand, show a comparative readiness to pass out of the dispersed phase as the dilution is increased. The behaviour of the potassium is the most pecuHar, for, as dilution goes on, the calculated concentration of this ion actually decreases, but as the decrease is slight it is probably due to experimental error, and it was treated as such by Bialascewicz himself. Thus, of the ions bound to the dispersed i | o-s phase, the cations sodium and potassium, and the anions of §? 0-3^, chlorine and (presumably) phosphate, are firmly attached, while the cations calcium and magnesium are not, and can easily be washed out. The high proportion of phosphorus in organic combination should be remembered here. Bialascewicz also pointed out that the partition coefficient or ratio followed with dilution a practically rectilinear course, so that some idea of the ratio in the natural undiluted yolk might be obtained by extrapolation. These figures so obtained are shown in Fig. 17, from which it may be deduced that the ions follow the order phosphorus, calcium, magnesium, chlorine, potassium, sodium, beginning with the one most of which is in the dispersed phase and ending with the one least of which is so distributed. Fig. 1 8 shows another aspect of the passage of ash from dispersed to continuous phase. In succeeding papers Bialascewicz extended these researches to the eggs of amphibia, some other fishes, Crustacea, molluscs, echinoderms and annelids. He reported that the intermicellar liquid varied much in its relative amount, accounting for from 20 to 63 per cent, of the whole ooplasm. From the data in Table 50, however, there does not seem to be a very close relation between the relative volume extrapolated 2 values for undiluted ooplasm 4 6 degree of dilution Fig. 17. 364 THE UNFERTILISED EGG AS A [PT. Ill 100 90 €1 -0% 70 M, Ca, 6 • — ■ «..;«  of continuous phase and the percentage dry weight of the system. Bia lascewicz's tables give the concentration in percentages of the principal ions in the intermiceliar liquid of different eggs, and these are conveniently summarised in Fig. 19, taken from his paper. From this it is obvious that all the eggs studied have about the same proportion of potassium, but that the other ions are rather variable. There is much more calcium, relatively, in the continuous phase of the yolk of the hen's egg than in that of any of the others except the crustacean Maia verrucosa. Similarly, there is more magnesium, relatively, in that of the frog than in any other egg. A very interesting comparison may be made between the distribution of ions in the continuous phase of the eggs and that in the serum of Table 50. Bialascewicz's figures. 0-1 0-2 0-3 Concentration of the bhree elements in the continuous phase (mgrnt per cc) Fig. 18. 0-4 Continuous phase cc. per cc. In% of vol. Egg ooplasm ooplasm % dry weight Scyllium canicula 0-83 17-0 — Salmo fontinalis 0-79 20-8 — Salmofario ... — 41-5 (Faur^-Fremiet & Garrault) Torpedo ocellata 0-41 59-0 — Acanthias vulgaris — — 47-3 (Zdarek) Arbacia pustulosa 0-82 17-8 — Paracentrotus lividus .. 0-79 20-7 22-6 (Wetzel) Rana temporaria o-6o 39-9 42-6 (Kolb, Terroine, etc.) Callus domes ticus 0-55 45-1 50-3 (Kojo) Sepia officinalis 0-50 50-0 47-3 (Wetzel) Maia verrucosa 0-37 63-2 43-6 (Wetzel) SECT, l] PHYSICO-CHEMICAL SYSTEM 365 the corresponding adult animals. Fig. 20, taken from Bialascewicz, shows that the potassium preponderates in the former and the sodium in the latter, while the other inorganic substances are more or less equally distributed. As there is no difference in electrolyte con 70 -S 604 to 50 40 30i 20 10

2 o CO I O o o ^ to I ^= Potassium ^ = Calciu = Sod'rum Fig. 19 ^ = Magnesium centration between the continuous ooplasm phases of fresh-water and marine animals, one must conclude that salts do not account for the properties possessed by the latter, and that crystalloidal organic compounds, such as taurine, urea and glycine, play an important part in keeping up a high osmotic pressure. Thus Bialascewicz found a concentration of 8-43 gm. per litre of urea in 366 THE UNFERTILISED EGG AS A [PT. Ill undeveloped Torpedo ocellata eggs, but none in those of Arbacia, Sepia, or Maia. Vladimirov occupied himself with the egg-white in the egg of the bird. In connection with other investigations which dealt with the Salmo fonblnalis PcGassium Sodium Magnesium Calcium Torpedo ocellaba Maia verrucosa Continuous phase of egg-yolk Serum of adult animal Fig. 20. water metabolism of the egg (see Section 6-4) he measured the electrical conductivity of the egg-white in the unfertilised ovum, using the Kohlrausch and Holborn apparatus, and obtaining a value of 7*6 X io~^. By the aid of a dialysis method he calculated the electrical conductivity of the intermicellar fluid of the egg-white, allowing for the disturbing effect of so large a concentration of pro SECT. I] PHYSICO-CHEMICAL SYSTEM 367 tein. The result came out to 10-4 x io~^. If this work were repeated for the yolk, interesting commentary on Bialascewicz's researches would be possible. It agrees with the earlier measurements of Bellini, who found the electrical resistance of the unincubated white to be Q. 1 8-8 ohms. Much further work on such properties of the yolks and egg-whites of a wide range of eggs is urgently needed, for they must obviously be of the greatest importance to the developing embryo. Such questions as the electrical conductivity of egg-cells and developing embryos are very relevant here, but must be left for consideration in Section 5. As a conclusion to this discussion of the chemical constitution of the egg, it may be admitted that great progress has been made in our knowledge with respect to it during the last fifty years. But to a discerning judgment, it remains none the less a matter for great surprise that in view of our comparative ignorance of the chemical architecture of the egg, we know as much as we do about the cominginto-being of the chemical architecture of the finished embryo. One further matter may be alluded to in this section. The composition of fossil eggs cannot be said to have much embryological interest, but it is hard to exclude a mention of them. The only analyses we have are those of ZoUer who worked on the fossil eggs of Chincha Island, off the Peruvian coast, where seagulls have been living and depositing guano from a very remote date. Zoller found that "time, which antiquates antiquities, and hath an art to make a dust of all things" had had that effect on these eggs and had reduced their water content to 14-4 per cent. There was no urea or uric acid present, although the protein had nearly all disappeared and had given rise to ammonium salts. There was no trace of fat or of carbohydrate, and the sulphur of the proteins had all turned into sulphate. Water % 14-4 Cholesterol 0-287 Phosphoric acid ... 0-045 Total nitrogen 945 Ammonia N 8-12 K 14-9 SOs 16-08 These figures make it only too clear that if palaeontology and biochemistry enter into closer relations than exist at present, it will not be by way of the chemical analysis of fossil eggs. More hopeful approaches will be found in Section 9-15.

Section 2 On Increase in Size and Weight

2-1. Introduction

We have so far been considering the unfertilised egg-cell and its reserves of nutrient material as a physico-chemical system, and we must now proceed to summarise critically what is known about the alteration the egg undergoes in passing into the state of the finished embryo. Subsequent sections will take up the chemical changes during this process in all their complexity, but first the apparently simple phenomena of change of weight must receive consideration. To this undertaking special difficulties are attached; for example, the act of birth or hatching itself, important though it is for the chemical embryologist as the term of his investigations, is yet purely arbitrarily and conventionally chosen as such, and, as far as the organism itself is concerned, may be relatively unimportant. The age at which birth takes place varies in different animals considerably, and may occur earlier or later in development, cutting across cycles of growth at almost any point. However, the study of growth in weight and alteration in shape, is an essential preliminary to the study of the chemistry of the embryo. I do not propose to spend any time in the discussion of definitions of growth. The actual data which we have concerning pre-natal growth will be found in Appendix i, where they have been placed in the hope that a collection of them will be of assistance to chemical embryologists. No previous assemblage of them has been made, and they are to be found scattered all through the literature. Biochemists have in the past been insufficiently careful to check their results on embryos against normal tables of weight, length, age, etc. The predecessors of this section are the chapters on growth in d'Arcy Thompson's Growth and Form and Faure-Fremiet's La Cinetique du Developpement. These authors gave a full criticism of the whole subject, but without special reference to the development of the embryo. Moreover, much has been done since they wrote, and their treatment differs in various ways from what follows here. PT. m, SECT. 2] ON INCREASE IN SIZE AND WEIGHT 369 It is obvious that the growth of an embryonic organism can be measured in many ways besides that of increasing weight. Its enlarging dimensions in various directions of space can be measured, or its volume, or the quantity of various constituent substances. More will later have to be said about the way in which these different quantities may be thought of as fitting in together and changing with age. But the simplest manner of representing growth will probably always remain the measurement of the increase in weight of the total mass, and it is this which is now to be considered. The relation of this factor to the age and the length of the foetus is a point of capital importance to the chemical embryologist in the knowledge of his material. It is true that the data are fragmentary enough, restricted as they are almost entirely to various mammals and the chick. 2'2. The Existing Data Cephalopods. Octopus. A remarkably complete set of data for the embryonic growth oi Sepia is given by Ranzi, and this is almost all we have as regards invertebrate development. Insects. Silkworm. Luciani & LoMonaco have studied the curve of growth through the successive moults in the larval condition, but, in spite of their work and of many other researches on the silkworm larva and ^gg, I cannot find any in which the increasing weight of the embryo itself has been measured prior to hatching. Fishes. Trout. Weighings of fish embryos have been exceedingly few in number, owing to the smallness of their size and the difficulty of separating them from the yolk. Kronfeld & Scheminzki, however, have made some estimations of the increase in weight of trout embryos, and their figures, together with those of Gray, are shown in Table i of Appendix i. 24 370 ON INCREASE IN SIZE [PT, m Amphibia. Frog. In the case of amphibia, where the cleavage in the egg is more or less inclusive of the yolk-laden portion, it is not possible to obtain data for the weight of the embryo itself, for, before hatching, although the protoplasm is constantly increasing at the expense of the yolk, the two elements cannot be separated, and therefore cannot be weighed in isolation. This appears in the figures of FaureFremiet & Dragoiu ; Schaper ; Davenport ; and Bialascewicz, and must always be taken into account when differences between species in water-content and other constants are under consideration, for much confusion may be caused by not distinguishing carefully between yolk plus embryo and embryo alone. Reptilia. Snake. Bohr's very few figures on Coluber natrix are all that are available. (Appendix i. Table 2.) Birds. Chick. It is on this animal, as might be expected, that the greater part of the work on embryonic growth has been done. Hasselbalch, in the course of his work on the respiration of the chick embryo, obtained a regular series for a race not given. These corresponded well enough with the earlier data of Falck (also from an unknown breed), which were the first to be published, appearing in 1857. Hasselbalch's curve is shown in Fig. 21, in which for the first time we see the usual ' ' embryo-placenta relation ' ' in the form of a weight of extra-embryonic structures larger than the embryo in the earliest stages, but soon falling below it.^ relation between the two as follows: 4-0 C3.0 2'0 1«0 O Membranes • Embryo Hasselbalch calculated the See also Fig. 521. SECT. 2] AND WEIGHT 371 Wt. of embryo + wt. of membranes Day Wt. of embryo 8 9 ID 15 16 1-917 1-652 1-613 I -108 1-128 17 I-IIO 18 1-090 Other sets of weight data have been reported by Lamson & Edmond, by Murray and by Needham, for White Leghorn chicks, and by LeBreton & Schaeffer for chicks of an unspecified race. These are all placed in Table 3 of Appendix i, where it will be seen that the general agreement between them is good. The values obtained by Murray; Byerly^ and Schmalhausen are probably the most accurate, for the conditions were very carefully controlled. Hanan's values are lower than all the others. It is unfortunate that Schmalhausen does not state what breed of hen was used in his experiments, though he does mention that it was not genetically pure. Some early measurements by Welcker & Brandt are not included in the table, for they do not appear to be trustworthy. Other measurements which are useful are those of Edwards, who has published a peaked curve showing the length of the primitive streak during the first 50 hours of incubation. ^ Schmalhausen's work on the growth of parts of the chick embryo will be dealt with later : he was preceded by Falck, who measured and weighed various organs but did not use enough material to make his figures valuable to modern workers. Mammals. [a) Mouse. In 1923 LeBreton & Schaeffer published figures for the embryonic growth of the mouse, but these were not very numerous. The only other work on this subject is that of McDowell, Allen & McDowell, probably the most accurate and satisfactory study of prenatal growth in any form that at present exists. Their figures are given in Table 4 of Appendix i, and the curve obtained from them in Fig. 22. This is drawn on arithlog paper, the ordinates in logarithmic ruling giving the actual weights in gm., the abscissae in arithmetical ruling giving the age and the number of the individuals. On each day the range of the individual unclassified weights is shown by a vertical line which is itself used as a base-line for the frequency distribution of the classified individual weights. The number of cases in the ^ See also Fischel and Leva. 24-2 372 ON INCREASE IN SIZE [PT. Ill distribution is shown by the distance to the right of the vertical baseHnes, and can be judged by the frequency-scale at the bottom of the WEIGHT GRAMS 1 -000 •1000 •0100 •00100 •00010 •00001 1 e^. J .u^^ \-y^ ^^^ > \^' '^ ^ > , zz Y - k/^ T V^Z ' />H V^ t / / / / / J 1/ .c;r.Al E np FRFniifNniF'=^ 1 I 1 1 1 1 1 1 1 1 1 1 1 1 20 40 60 1 9 10 n 12 13 14 15 CONCEPTION AGE Fig. 22. 16 17 18 19 chart. The means, weighted by the number of individuals in each litter, are shown as dots on the vertical base-lines, and it is through these, of course, that the "normal curve" would be drawn. The continuous curve in the graph is one drawn to a formula which SECT. 2] AND WEIGHT 373 Foetus of albino rab will be discussed later, in Section 2-4 (p. 393). The lay-out so made reveals several interesting features; it appears, for example, that there are always individuals on a given day which are equal in weight to the mode of the day before. McDowell, Allen & McDowell consider that this is evidence of a possible delay of as much as 24 hours between copulation and fertilisation, but, whether this is so or not, it certainly equates with exactly similar variations found in the chick both in the early stages (primitive streak) and in the later ones of organ-growth. Further, the modes and means are generally close together, though less so at the beginning of development than at the end, and the latter do not approximate to a straight line. A glance at the graph also shows that the highest individual weights on each day tend to form a curve parallel with that of the means throughout development. (b) Rat. Donaldson's comprehensive monograph of 191 5 includes a discussion of the growth of the rat embryo, but much less work has been done on this animal than upon man, for in the latter 5 case the ad hoc labours of obstetricians have often provided much valuable material for the biologist. However, Stotsenberg's work gave a good account of the matter, and his figures are reproduced in Table 5 of Appendix i, and in Fig. 23. They begin from the 13th day after insemination, before which weighing is difficult, and they continue until birth, which takes place at the 22nd day. This prenatal period would appear to be one complete growth-cycle, if we may judge from the work of Donaldson, Dunn & Watson on the post-natal growth of the rat. Huber has studied the growth of the rat embryo in its earliest stages prior to fixation to the uterine wall. He states that the eggcell of the rat approaches the uterine end of the oviduct while in the two-cell stage, segmentation being slow and proceeding as the transit takes place. Fig. 24, reproduced from his monograph, is a 16 17 18 Fig. 23. 374 ON INCREASE IN SIZE [PT. Ill photograph of a model of the oviduct with its contained eggs. By reconstruction methods at a magnification of looo diameters of the ova Huber was able to determine the volume changes during segmentation as follows: Age bays ^ , Hours Stage Average vol. (cubic mm.) I Pronuclear 0-000156 2 o 2-cell 0-000162 3 I 4-cell 0-000173 3 17 8-cell 0-000184 1 1 -cell 0-000210 There would, therefore, appear to be a certain increase in volume during these very early stages, but as the specific gravity changes are not known it is difficult to understand what it may imply. There is at present a great gap in our knowledge of the embryonic growth Fig. 24. of the rat between the early point at which Ruber's studies end and the later one at which those of Stotsenberg begin. Huber himself suggested that the slow development of the ovum of the rat during its passage down the oviduct was best accounted for by the lack of any food-supply for an alecithic egg until fixation to the uterine wall had taken place. As the whole embryonic period of the rat is only 22 days, it is of great interest that the first four days should involve hardly any increase in size. This fact renders of no significance the calculated weights of rat foetuses given by Donaldson, Dunn & SECT. 2] AND WEIGHT 375 Watson in their earlier paper, for, in assuming that embryonic growth in the rat followed a quite similar course to that taking place in man and the rabbit, they did not allow for the long time taken for the rat egg to pass through the oviduct after fertilisation. Thus they arrived at the result that the rat embryo of 15 days should probably weigh 2-6 10 gm., whereas by direct measurement Stotsenberg found that it only weighs o- 1 68 gm. Their calculated figures are consequently not included in Appendix i. (c) Guinea-pig. The most usually quoted work on the embryonic growth of this animal used to be that of Read, who used a very indirect method of measuring it. He weighed the pregnant female every day between insemination and birth and then each foetus with its membranes and fluids, from which data, assuming that growth had taken place regularly, the weight of one embryo could be calculated. He concluded that the guinea-pig passes through two growthcycles during its intra-uterine life. But no satisfactory conclusion can really be drawn from such figures, subject as they are to all kinds of complicating factors, and, like the earlier ones of Minot on the guineapig, obtained in the same indirect way, they are better discarded. It is needless to point out that differences in the weight of mother + embryo due to defaecation, filling of caecum, etc., may amount to grams, while the weight of the embryo is still only milligrams. In the absence of any other figures, they had their importance, but in 1920 Draper made a complete study of the embryonic growth of the guinea-pig. Together with the few fragmentary (but direct) figures of Hensen, and the careful work of • 90 — 80 • / • / / • 70 " / • 60 - E .7 ( 50 - C tj 40 .«; •/ A * • 30 ~ • / • / 20 J 10 ) / 1 • 11 days, 1 1 10 20 30 40 50 60 70 Fig. 25. 376 ON INCREASE IN SIZE [PT. Ill Ibsen, and Ibsen & Ibsen, Draper's figures form the standard series, and are shown in Tables 6 and 7 of Appendix i. As is generally known, the guinea-pig differs from most other mammals in being born much later in its life-span than is usual, so that its lactation period is exceedingly short and it is able to eat green food a very few days after its birth. This is reflected in its gestation time which is relatively long. 20 '0 CO / / 15«0 —~ -0/ CO ^ E ® / a '^ / %. / a> / 10*0 — c ^ a:> K®/ ^ • uv- of 'membranes ? • 5«0 yA IX, Age ab which wbs. of embryos &, _>*x'^ membranes are equal O'O -1--^ i^;;i^ Aqe in days 15 20 25 30 35 40 Fig. 26. 45 50 55 60 65 During the 64 days of its development in utero, the guinea-pig increases its weight to about 85 gm. and its length to 10 cm. This process is shown in Fig. 25 taken from the figures of Draper. In Fig. 26, which gives an enlarged view of the lower part of the growthcurve, the increase in weight of the placenta, the membranes, and adnexa, together with the amniotic and allantoic fluid, is also shown. The extra-embryonic structures reach a more or less constant weight about two-thirds of the way through development, but, as can be seen from Table 5 of Appendix i, the values from which this curve was drawn are very divergent. In comparing the growth of the SECT. 2] AND WEIGHT 377 embryo with the growth of the membranes, it is interesting to see that for the first month the latter weigh much more than the former, after which, for a certain period, they grow together at the same rate. But soon the curves diverge, and the membranes hardly grow any more, while the embryo continues to increase greatly in size. Evidently when the membranes and placenta have reached a sufficient size to meet the utmost further demands of the embryo they grow no more. There can be little doubt that the size of the placenta exercises an influence on the growth of the embryo, and is of the highest importance from the point of view of embryonic nutrition. The amniotic liquid bears the same relationship as regards weight to the embryo as do the placenta and the membranes. 100 Fig. 27. Fig. 27 shows Draper's curve for the length of the embryonic guinea-pig. Ibsen's work led to much the same conclusions as regards the relations between embryo and adnexa as that of Draper. Ibsen found that the number of foetuses in the uterus exerted an effect on the growth-rate of each one, thus the larger the litter the slower the rate of growth of the individual foetus. The early growth of the placenta is more rapid than that of the foetus, but they reach the same weight on the 25th day, after which the foetus outstrips the placenta very soon. Placental weight and the weight of the membranes towards the end of pregnancy are closely correlated with uterine crowding, but this is not the case with the decidua basalis, which corresponds to the maternal part of the placenta. Minot considered that the amniotic fluid of the cow and of man decreased in 378 ON INCREASE IN SIZE [PT. Ill amount after the middle of pregnancy, but this was not found to be the case by Ibsen for the guinea-pig. Ibsen constructed from his experimental data the interesting diagram shown in Fig. 28, which shows the percentage of the whole system occupied by embryo, placenta, decidua basalis, and amniotic fluid from the 20th day onwards. The embryo does not rise in per cent, after the 55th day, the placenta remains very much the same all through, the decidua basalis is much smaller relatively at the end than at the beginning and so is the amniotic fluid. Up to the 50th day Ibsen found no correlation between foetus-weight and placenta-weight, but afterwards there is undoubtedly such a correlation, evidently due to crowding. [d) Rabbit. Much less work has been done on this form than might have been expected from its easy availability, but the figures of Chaine (the standard ones) are given in Table 8 of Appendix i, together with some early fragmentary ones of Fehling. Friedenthal also gives a few data which are shown in Table 9. {e) Dog. Liesenfeld, Dahmen & Junkersdorf made a thorough study of (unfortunately only 5!) dog foetuses. SECT. 2] AND WEIGHT 379 (/) Sheep. As early as 1847 Gurlt made a study of the increase in length of the foetus of the sheep, but Colin is the only investigator who has ever determined the growth in weight (see Fig. 29). Gurlt's figures, which are quite regular, are given in Tables 10 and II of Appendix i. FaureFremiet & Dragoiu, in the course of their extended work on the growth and chemical development of the embryonic lung in the sheep, made measurements of the growth of that organ, but did not give any data on the weights of their foetuses as a whole, a very unfortunate omission in view of the incompleteness of the literature on this subject. [g) Pig. The only extensive figures in existence for the embryonic growth of the pig are due to the careful work of Lowrey and of Warwick. These are given in Tables 12 and 14 of Appendix i. Lowrey's results will again be referred to in connection with the growth of individual organs and parts in the embryo. LeBreton & Schaeffer also measured and weighed a certain number of foetuses in the course of their classical work on the behaviour of the chemical nucleo-plasmatic ratio during 60 100 Days, Sheep (Colin) Fig. 29. embryonic development. Their figures are shown in Table 1 3 of i Appendix i. \ (Ji) Cow. The embryonic \ growth of the cow has been \ investigated by several workers whose results are shown in ' Table 15 of Appendix i. Fig. 30, constructed from Franck and Hammond, should be compared with Fig. 28 for the guinea-pig. {i) Man. The embryonic growth of man has been much studied, and many thousands of embryos have been weighed. His's studies have been the principal means of fixing the relation age/length, Months, Cow (Hammond). Fig. 30. 38o ON INCREASE IN SIZE [PT. Ill and Balthazard also gives figures for this, which will be found in Appendix i (Table i6). The earlier workers, Ahlfeld; Fehling; Hennig; Legou; Faucon; and Michaelis all obtained valuable data, but it was not until 1909 that a critical examination of them was made by Jackson who analysed the figures of his predecessors, and added a large number of new ones. His results gave a continuous curve from the earliest stages till birth, which agreed with the majority of the other investigators, but not perhaps with Hennig's curve (he gave no figures), which showed a very distinct slackening of growth about the sixth month, after which the same rate was resumed. If this phenomenon is real, it may possibly be associated, as Donaldson has suggested, with a transition from one growth-cycle to another, at the end of the sixth month, when the absolute weight begins to rise so rapidly. On the other hand, the mass of data which Quetelet and others after him have analysed regarding the growth of man throughout life, would seem to show that there are three growth-cycles only, one pre-natal one, one with its maximum at 5*5 years, and the third with its maximum at 16 years. Vignes' S-shaped curve for human embryonic growth is shown in Fig. 31. Bujard in 19 14 made a geometrical analysis of the early stages of the human embryo. Jackson measured the volume and weight of all the specimens in his own collection, and for the early stages also the volumes of the His-Ziegler models. His figures are given in Tables 17 and 18 of Appendix i, and the curve which he constructed from his own data as well as those of previous investigators is reproduced in Fig. 32. The 1 6th table of the appendix shows the volumes of the HisZiegler models, and demonstrates that the human embryo, like all others, is much exceeded in size by the yolk-sac during the earliest 3500 3398<jyP3405 31410^88 3000 im(p 2000 n E TO i. - C 1000 a3 '© mcf 1 500 100 JH no. 305 7 1 _. 1 260 280 300 1 1 i > f 1 1 1 ■ 3 4 5 Months Fig. 31 .270 290 310 Days SECT. 2] AND WEIGHT 381 stages of development. Jackson, who adopted the Minot method of measuring the growth-rate, concluded that the rate was 9999 per a) E o > E o <4 © E j3 CO 0) (0 E o^ c X) j: 3250 3000 2750 2500 2250 2000 1750 .1500 1250 1000 750 500 250 l: •

\ 1 i <l / / !j 1. 1 AhlfelcL's dauta. ■ehling's da-bet Jaxjkson's data, L.egoas data, Michael is' data 1 i 1 * It 2. 3. i i '/• 4 5. L 1 ! 1 1 1 1 ! 1 J / i 1 / / / / / 1 t I 1 i / 1 1 / / / / / / 2// 7 / 1 ' / / -4 '•• / ^ j^' L^. ^ ^ ' • 50 75 1( 125 150 175 200 225 250 275 Age in Days Fig. 32. cent, for the first month, 74 for the second month, and 11 for the third month. This was in general agreement with the point of view taken by Miihlmann, and Jackson emphasised it further by showing that, if the weight of the embryonic membranes and fluids was taken 382 ON INCREASE IN SIZE [PT. Ill into account, the growth-rate for the first month was 574,999 per cent. What meaning can be attached to the enormous growth-rate figures which always appear when the Minot method is used for very young embryos must later be discussed. Fig. 32, which collects together the data of many observers, shows a considerable measure of unanimity between them. Ahlfeld's figures are the only ones which show serious divergence, and they were not taken into account by Jackson in his preparation of the "normal curve". Fig. 32 shows also by points the volume of the embryo at the different stages, but it does not differ much in value from the weight in grams. The specific gravity of the foetus does not, according to Jackson, remain precisely the same throughout develop SECT. 2] AND WEIGHT 383 merit, but changes from very slightly above i-o in the early stages to 1-05 in the later ones. Probably this is associated with the progressive loss of water as the embryo develops. One of the first to investigate quantitatively the growth of the human embryo was Boyd in 1861, who studied the weights of all the principal organs in embryos from 8*5 to 85 oz. He did not give figures for individual embryos from which a curve could be constructed, but simply divided them into large groups such as "prematurely-born", etc. Legou's data, already referred to, were worked over again in 1903 by Loisel. Zangemeister, more recently, has published figures for human embryonic growth — these are shown in Table 17 of Appendix i. Other data for embryonic growth in man will be found in the papers of Fesser ; Toldt ; Meyer ; Heuser ; Bedu ; Sombret ; Arnoljevic ; Stratz; Borri; Corrado; Balthazard & Dervieux; Ecker; Hamy; Kolliker; Cruickshank & Miller; Browne; and Friedenthal. Scammon & Calkins, who have made a great many measurements in recent years, have constructed a three-dimensional isometric projection, from which the height, weight and age of a human embryo may be read off if any one of them is known (see Fig. 33). The best recent paper on the whole subject is that of Streeter. Sandiford has shown that the weights and surface areas of foetuses fall on straight lines when plotted on double log paper. For further information on surface growth see Scammon & Klein. (j) Whale. Some information on the embryonic growth of the whale is contained in the papers of Harmer; Risting; Hinton; and Mcintosh & Wheeler, but it mostly concerns increase in length. 2-3. The General Nature of Embryonic Growth We may now turn to the theoretical aspect of the matter in the attempt to find out what interpretation can be placed upon them. We may in the first place take as a simple example of an embryonic growth-curve the work on the growth of the chick (White Leghorn) of H. A. Murray. Table 51 shows, firstly, the actual weights of the embryos on each successive day, secondly, the amount gained in each such 24-hour period, i.e. the amount of substance actually added on to itself by the embryo during the lapse of the time in question. This is known as the daily increment. In the next column the averages of the daily increments are placed, and these figures, known as the 384 ON INCREASE IN SIZE [pt. hi mid-increments, represent for each point which begins one period and ends another how much substance the embryo is adding on to itself between the times (a) half-way through the last period, and (b) half-way through the period to come. In other words, the midincrements convert the daily increments into terms of the points instead of the spaces between the points. If now the mid-increments Table 5 1 . Growth of the chick embryo ( White Leghorn) . H. A. Murray's figures. Percentage Daily Mid growth-rate .Age Wet weight Dry weight increment increment of dry in days (mg-) (mg.) (dry weight) (dry weight) substance 5 221 "•75 "•85 6 423 23-6 19-4 15-7 66-5 7 735 43-0 30-8 25-1 58-4 8 1,189 73-6 44-3 37-5 50-8 9 1,817 ii8-i 68-2 56-2 47-5 10 2,661 186-3 102-5 85-3 45-7 II 3,750 288-8 160-7 131-6 45-6 12 5.105 449-5 241-0 200-8 44.7 13 6,839 690-5 409-4 325-2 47-1 14 8,974 1099-0 575-0 492-0 44.7 15 1 1 ,460 1674-0 686-0 630-0 37-6 16 14,390 2360-0 730-0 708-0 30-0 17 17,950 3090-0 797-0 763-0 24-7 18 22,030 3887-0 832-0 814-0 20-9 19 26,670 4719-0 are expressed as percentages of their actual weights of embryo at their corresponding points in time, the last column or percentage growth-rate is obtained. This last calculation is, as will be seen, the only one so far made in the table which is open to serious criticism, and it is associated with the name of C. S. Minot, who was the first to propose it as a satisfactory measure of the growth-rate of an organism. When these figures are plotted the curves shown in Fig. 34 appear. The actual growth of the embryo expressed in terms of its SECT. 2] AND WEIGHT 385 weight at any given moment gives a curve which rises steadily till the observations cease without betraying much sign of any slackening. But the increment curves, on the other hand, show an unmistakable S-shape which is due to the fact that, for the earlier periods, the — o AbsoLate weight gms. wet • >> )> j> dry — Dajly incremenbl ^ « Mid » r^^ Fig. 34 weight gained each day is very little more on one day than the gain on the previous day, while, towards the middle of development, the daily increments and the mid-increments vary much more, each one being considerably higher than the one before. On the other hand, when the end of development is approaching, the increments each day, though far higher in absolute amount than those which were made in the early stages, do not differ so materially from one another, with NEI 25 386 ON INCREASE IN SIZE [PT. Ill the result that the curve slackens off and enters a slowly-rising phase again. It is possible, of course, to calculate the average daily increment for the whole developmental period (see Table 52), and the figures so obtained have been made the basis of a comparison of animals by Friedenthal. The fourth curve, that of the percentage growth-rate, Minot's curve, as it may be called, begins at a high level and continually descends, although in this instance there is a slight kink on it midway through development, which may for the moment be disregarded. All Minot curves begin at a high level and descend as development proceeds. Now, in many cases, it may happen that not only the increments but also the whole growthprocess itself slackens off towards the end of the period taken, in which circumstance the curve relating weight of animal at any given time to age will also have an S-shaped form. It has not escaped the perspicuity of those who have considered these phenomena that this S-shaped curve has a resemblance to the S-shaped curve of an autocatalysed monomolecular reaction, and this likeness will shortly be taken up at length. Table 52. Average daily increments. Friedenthal's figures. gm. gm. Mouse o-o8 Pig 14 Rat 0-24 Man 15 Ermine ... 03 Sheep ... 26 German marmot 1-2 Seal 30 Musk 1-5 Ass 53 Guinea-pig 2-0 Rhinoceros 90 Wolf 4 Stag ... 100 Puma 5-4 Horse ... 200 Bear 7-0 Hippopotamus . . 200 Lion 10 Camel ... 400 Roedeer... II Elephant 400 A more complicated example of the various types of growth-curves is afforded by Fig. 35 taken from Brody. It shows the growth throughout the life-span of the albino rat. The curve passing through the circles shows the course of growth ; it is, in fact, the weight of the whole animal at any given moment plotted against the age at that moment. The strongly indented curve, passing through the crosses, is the line showing increment in unit time. In the data of Murray for the embryo chick the absolute growth-curve rises steadily. SECT. 2] AND WEIGHT 387 and has no slackening off or self-inhibitory phase; the increment curve is therefore singly sigmoid. But here, when the absolute growth-curve is itself sigmoid, the increment curve is symmetrically sigmoid, rising to a maximum and then falhng away again to zero during the second phase. Finally, the corresponding Minot curve is shown by the line joining the triangles, and, as usual, it declines throughout growth from an initially very high value. n 70 30 «25 n) u CO a»20 m i. OU 60 <0 2 40 ■^30 0.20 gms. 200 cent day DaysO 10 20 30 40 50 60 70 80 90 100 HO 120 130 140 150 160 lyO^lB0 190 200 210 220 o 8 18 28 38 48 58 68 78 88 98 108 118 128138148 158 168 178 188198 % I Age o U Fig- 35 There are other ways, however, in which the subject of embryonic growth-curves can be introduced. Ostwald's classical work on growth in metazoa, which appeared in W. Roux's "Vortrage" series in 1908, laid great emphasis on the value of knowing the precise route through weight taken by an organism on its way from egg-cell to finished embryo. In Fig. 36, taken from his monograph, several different curves are shown relating time to weight. At the time A, at hatching or birth, for instance, the weight of the 25-2 388 ON INCREASE IN SIZE [PT. Ill organism is exactly the same in all four cases, but the manner in which the increase in weight has taken place is in the four cases profoundly different. It is certainly quite clear that the chemical embryologist, engaged in the attempt to understand the processes which contribute to the final result, must pay detailed attention to the path by which this final result is arrived at. The four different curves in Ostwald's figure would imply four very different sets of conditions within the developing embryo. An embryo which grew according to Curve I would change very rapidly in the beginning, and afterwards change progressively less rapidly as the curve became asymptotic. The reverse of this process would happen if the embryo grew according to Curve ii, for there the process continually increases in rapidity, and is proceeding at its fastest when the point A is reached. Curve iii, on the other hand, being S-shaped, would seem to indicate the presence of an autocatalytic process, for at first the growth proceeds faster and faster, but later on, after the point of symmetry of the curve has been reached and passed, slower and slower. Several such S-shaped curves superimposed on one another make up Curve iv. As far as is known, no growth takes place in the manner represented by Curve i, but rather in that of the other three curves, though our present knowledge does not enable us to say definitely which, except in certain cases. Ostwald's monograph should be referred to by those who wish to see how he continued the discussion of growth-curves, for it is probably the best presentation of the subject, and it was certainly written from a much less doctrinaire point of view than most of its successors. The general interpretations of embryonic growth-curves may be divided into several classes. They depend more than anything else, as will be clearly seen, upon how the facts are expressed. One way of expressing them led Minot to his "laws of cytomorphosis", another led Robertson to his "autocatalytic master-reaction", SECT. 2] AND WEIGHT 389 and, more recently, still other ways have been devised. The unprejudiced investigator cannot avoid a considerable measure of scepticism in considering the claims of one way of expressing the facts over another. 2-4. The Empirical Formulae We may first direct our attention to those presentations of the facts which do not carry with them any theoretical superstructure, but aim simply at describing the data in as short a manner as possible. The first of these "empirical formulae" was that of Roberts, who in 1906 pointed out that the growth of the human foetus could be regarded as nearly proportional to the cube of the age ; thus, if the weight in grams is W and the age in days T, the formula would be W — T^. But this was only very approximate, and the curve it gave did not fit the curve drawn through the experimental data with any accuracy. Roberts, indeed, stated that his formula gave results correct to "within an ounce at the third month". "Since the weight of an embryo of the third month," was Meyer's remark, "according to the best available evidence, is considerably less than an ounce, the accuracy of Roberts' method must be fully apparent without further comment." Tuttle next introduced an equation in which arbitrary constants were introduced, thus W = 50 {T — 2)^. Later still, Jackson, whose work on the human embryo has already been mentioned, proposed the formula : where W is the weight in grams and T the age in days. This fitted the experimental points much better than the formulae of Roberts and Tuttle, but was still rather deficient, especially in the very early period and the very late period. Henry & Bastien also proposed x^ + 2^xj> — 3q>'2 — i62y = o, where x = months andj = kilos. Duvoir has reviewed the other more or less practical rules which have from time to time been proposed, such as Casper's rule that, from the fifth month onwards, the age of an embryo in man can be found by dividing the height in centimetres by 5. Balthazard & Dervieux 390 ON INCREASE IN SIZE [pt. iii altered this formula to 5-6. Again, Mall's rule states that the number of days embryo age is equal to the square root of the foetal total length in centimetres x 100. Balthazard & Dervieux have also evolved formulae relating foetal age to the length of the limb-bones, e.g.: L = femur length x 5-6 + 8 cm. L = humerus length x 6-5 + 8 cm. L = tibia length x 6-5 + 8 cm. The use of empirical formulae in the description of human foetal growth has been carried to its greatest refinement in the work of Scammon & Calkins, whose formula, n- 2-5/, L2 holds with great exactitude from the third month onwards. Another of their formulae, T = 2-134 X o-iZ X o-ooiiL^, holds with rather less exactitude from 2-5 foetal months onwards. In both these cases, T is the menstrual age in lunar months, L the total or crown-heel length of the dead body in centimetres. They also found that W= (o-26L)3-i'>8 + 4-6, 3 108 , or Z, = 3-846 VW - 4-6, where W is the weight of the dead body. From these equations, it follows that 3 108/ 15S4 / T= 2-134 + 0-3846 VW — 4-6 + 0-01627 VW — 4-6, or T= 3-0 + ^•04.gVw — 0-012, orW= 0-561 — 0-366 T X 0-061 T^. The formula of Donaldson, Dunn «& Watson, for the post-natal growth of the white rat- up to 80 days, W = a + bT + cT^, and after 80 days, W = a log T — bT — c, was of the same type as the other equations mentioned, but it had the additional refinement of including constants, a, b and c, which were variable according to sex and other factors. Murray, in his study of the chemistry of embryonic development, SECT. 2] AND WEIGHT 391 found that his series of chick embryo-weights could be accurately described by the equation: T3-6 1-496' or W=KT^-\ where K = o-668. This was not unlike the Balthazard-Dervieux formula for the human embryo: T= 19-4 X \^W. 23 Diojs 5 10 U t2 13 14 Incub&lion 2052 Fig- 37 Murray plotted the log. wet weight against the incubation age, and obtained a curve concave to the abscissa (see Fig. 37) corresponding to the curve which McDowell, Allen & McDowell got for the mouse embryo (see Fig. 22). He also found that the relation of log. weight to log. age was a straight line as far as his series of weighings went, and showed that the weighings of Hasselbalch and of Lamson & Edmond fell on the same straight line (Fig. 38). Murray's formula gave very good agreement with his figures, but these did not extend further back than the fifth day of incubation. When, later, Needham 392 ON INCREASE IN SIZE [pT. m and Schmalhausen made weighings of embryos between the second and the sixth day of incubation, it was found that Murray's formula did not hold for these earlier stages. Fig. 39 taken from the paper of McDowell, Allen & McDowell, illustrates this point. The broken line is drawn to Murray's formula, and the dotted line is an extrapolation of his curve which I made on the assumption that embryos grew at the same rate before 5 days as between 5 and 7 days, i.e. exponentially. The circles with dotted centres are the values experimentally obtained by me, the dots are those obtained by Schmalhausen, and the cross within a circle is Murray's earliest figure. 4.6 4.2 >.? 3 J.4 Of 3 2.6 2.2 fi^ ft. %^ H » t gjp ° V^ t A ^ 8 \^

^ " ^ X LSkiDSon i Ldmond (a^vep^oe of lo embrvos) • H&ss2lbailch (sln^l? ojeigninos) Single cuci§hin§s l^ken in the course of other experiments in this series ■ »^ ^ ^ -" ' 1 ^ 1 0.65 0.70 0.75 0.85 0.90 . 0.95 1.00 1J)5 L05 incubation a^e (dsajs) Fig. 38. U5 1.20 1.25 UO Murray's formula gives a line consistently above the experimental points for this early period, and the exponential extrapolation is quite at variance with them. But McDowell, Allen & McDowell evolved an equation which fitted these early points (the solid line) as well as all the later ones, as follows : log W = 3-436 log {10 (r- 0-5)} + 7-626. This new equation was based on an entirely different viewpoint from that of previous workers. McDowell and his collaborators regarded not "conception age" but "embryo age" as the right zero hour to take in growth calculations. It had always been assumed previously that conception or even insemination was the right starting-point, and Brody & Ragsdale and Brody based their method for finding age-equivalence in animals on this view, while Friedenthal SECT. 2] AND WEIGHT 393 had shown a similarity in relative growth-rates by plotting the log. weights against log. conception ages. McDowell and his collaborators, on the other hand, suggested that the time of growth ought rather to be calculated from the time at which the embryo first begins to have an axis, i.e. from the primitive-streak stage. Thus the major differences between the pre-natal growth of the guinea-pig, WEIGHT . INMGS. ■200 -180 •160 •140 -120 -100 • 80 ■ 60 ■ 40 - 20 DATA OF • SCHMALHAUSEN ONEEDHAM e MURRAY INCUBATION 1 AGE FREQUENCIES ©85 077 ©88 © ©83 77 ffi200 Fig. 39 the mouse, and the chick would be accounted for by the varying times taken to get through the preliminary work of arrangement and organisation. Processes such as gastrulation, according to this view, would involve a law of growth so different from the later axial type that no formula should be expected to cover the two. We have already seen in the case of the rat's egg that some considerable time may elapse between the time of fertilisation and that of fixation to the uterine wall, during which the supply of nutrient material may be 394 ON INCREASE IN SIZE [pt. iii very different from what afterwards obtains during embryonic development. There is, therefore, much justification for the view of McDowell and his collaborators. Brody himself had come nearly to this position without recognising it, for, in a paper published in 1923, he had pointed out that, during a short period in the early stages of growth (or regeneration) the apparent observed speed seems to be slower than would be expected. Thus the curve of the fitted equation cuts the time axis not at zero, the beginning of growth, but a little later. He advanced the explanation that Durbin had already applied to the initial slow phase in the regeneration of tadpole tails, namely, that a "cap of embryonic cells" was first formed, following in its growth quite different laws from the subsequent process as a whole. "It is suggested", said Brody, "that the apparent initial slow phase of growth of the individual from the fertilised egg is due to a similar qualitative growth." (Estimated weights of eggs are shown in Table 53.) McDowell and his collaborators proceeded to show that a similar formula would fit very well the curves of growth for the guinea-pig (Draper; Hensen; Ibsen & Ibsen), the mouse (McDowell, Allen & McDowell) and the chick (Murray; Needham; and Schmalhausen). For the mouse it was : log W = 3-649 log {10 (^ - 7-2)} + 8-6587; and for the guinea-pig it was : log W = 3-987 log {10 {t - 12)} + 5-1839; The significant thing about these empirical formulae is the deduction of a certain time in each case from the conception age, thus 7-2 days in the case of the mouse, 1 2 days in the case of the guinea-pig, and 0*5 day in the case of the chick (Allen & McDowell). The evidence on which McDowell and his collaborators rested their case for this shortening of the development time was drawn from various sources ; thus, from their own histological observations they found that the primitive streak in the mouse embryo appeared about the 7th day of development, for 6-day embryos show no mesoderm, while 7-day ones do, and usually the primitive groove as well. Sobotta's work is in agreement with this estimate. Their estimate for 12 days as the time taken for the embryo guinea-pig to reach the primitive-groove stage was based on the generally accepted work of BischoflT and Lieberkiihn, while, for the chick, Duval, whose illustrations are the SECT. 2] AND WEIGHT 395 usual "normaltafeln", shows the first appearance of the primitive groove at the loth hour of development, an assessment which is agreed to by Jenkinson and by Foster & Balfour (12th hour). Table 53. Probable dimensions of egg-cells. Weight Diameter in grams in fi (Friedenthal) (Hartmann) Gyclostomata Amphioxus ... o-oooooi — Lamprey ... ... 0-004 — Pisces Sturgeon ... 0*004 — Pike o-ooi — Amphibia Frog ... 0-004 — Reptilia Crocodile ... 40-0 — AVES Hen 45-0 — Aepyornis I20000-0 — Mammalia Monotremata Platypus — 2500 Spiny anteater ... o-i 3000 Marsupialia Dasyurus — 240 Opossum O-OOI 150 Edentata Armadillo — 80 Cetacea Whale — 140 Insectivora Mole — 125 Hedgehog — 100 Rodentia Mouse ... O-OOOOOOI 72 Rat — 72 Guinea-pig O-OOOOOI 80 Lagomorpha Rabbit 0-000003 125 Carnivora Dog 0-000003 140 Cat — 125 Ferret — 120 Ungulata Horse ... — 135 Sheep ... — 120 Goat — 140 Pig 0-000003 130 Cheiroptera Bat — 100 Lemurs Tarsius ... — 90 Primates Gibbon ... — "5 Macacus — 115 Gorilla — 135 Man 0-000004 135 "The general course of pre-natal growth in the mouse, the guineapig and the chick, can be expressed by straight lines relating the logarithms of the weight and the age only when age is counted from the beginning of the embryo proper." Such is the conclusion of McDowell and his collaborators, and, though it may seem barren 396 ON INCREASE IN SIZE [PT. Ill in theoretical results, it is nevertheless based on sounder considerations than the more ambitious ones of other workers*. It is probably legitimate to assume that the laws of growth before the formation of the embryonic axis are very different from what they are afterwards. It is also legitimate to assume that the differences between the velocity constants in the three formulae are due to the varying amount of organisation which has to go on in each case before the ^ formation of the primitive < groove. Fig. 40 shows the o straight-line relationships ? found to hold by McDowell, ^ Allen & McDowell in the 9 case of the guinea-pig, mouse ^ and chick, and Fig. 41 gives further examples, from which further variants of McDowell's formulae could easily be calculated. Clearly in embryonic growth log. weight is always proportional to log. time. With respect to Fig. 39, in which the weights of very young chick embryos are given, it should be noted that the discrepancy would naturally be expected to occur only in the early stages, for in the later ones the difference between conception age and embryo age would be a smaller percentage of the total. The

But see p. 427. 100 • n r t I 7 1 1 o- ono / I 9 • / 7 > '-A I J \ // / ^ // /i! J '■il P ■h 7 VI' •IOC

. L / i rT ■// // / •OK s // ' / i / / 1 f UJ ho J h h^ '§ c HICK •GO •^ // ^ MURF ?AY r/ NEEDHAM c • schmalhausem; .1 •00010 I GL INFA PIG / 1 BSEN ♦ DRAPER + V.HENSEN •00001 1 2 4 10 20 4060100 200 400 600 EMBRYO AGE IN }\J"^ OF DAY Fig. 40. SECT. 2] AND WEIGHT 397 averages for the early embryos reveal the difference by bending away from the lines drawn on the basis of incubation or conception age. Schmalhausen, continuing earlier work on the growth of bacilli and protozoa, has also put forward empirical formulae for the embryonic growth of the chick, but his equation \^W= T, while fundamentally the same as that of Murray, has no velocity constant. Fig. 42, taken from Schmalhausen's paper, shows that the cube root of the weight plotted against the age only gives an approximately straight line. Schmalhausen has included in the same figure the curves obtained by other methods; thus curve P' is the Minot (percentage growth-rate) curve for the wet weight, and Ps' for dry weight, while the curve marked log o- ip is the log. weight plotted against the age. As we have already seen, in the case of McDowell's figures for the mouse, and Murray's figures for the chick, this value always gives a curve rising concave to the abscissa. The curves P and Ps in Fig. 42 represent the absolute wet and dry weights respectively. Other empirical formulae have been proposed for growth-processes from time to time. Embryonic growth can be expressed roughly by exponential curves ; thus: W = wp\ where W is the mass of the embryo at time t, w the original mass, and p a constant. Thus the equation of an exponential curve is one in which the power is always changing. Janisch has given a discussion of the use of the exponential curve in all departments of biology, and in it he shows how important this relation is in growth phenomena. The "law of compound interest "however, put forward byBlackman in 1919 for the growth of plants, and which has been shown by Luyet to be a special form of the exponential relation W= w {i + ry, has not so far been of any assistance in describing embryonic growth. Another form of the exponential curve, the arithmetical progression method, which gives the equation log W - At, 398 ON INCREASE IN SIZE [PT. Ill where A is 3. characteristic constant, was used by Faure-Fremiet in 1922 for describing the growth of a Vorticella colony, but the con 100 CHICK Hasselbalch 1900 O Series (a) D « (h) Bohr& Hasselbalch 1900 • LamsonS^Edmond 1914 Ijin 1917 O LeBrebon8(Schaeffer1923 B Murray 1926 Schmalhausen 1926 ^ Needham 1927 RAT Stobsenberg 1915 MOUSE ® McDowell etc. 1927 <§> LeBreton&Schaeffer1923 GUINEA-PIG O Draper 1920 SHEEP ® Colin 1888 PIG O LeBrebon&.Scliaeffer1923 O Warwick 1928 TROUT X Scheminski 1922 RABBIT O Chaine 1911 « Lochhead&,Cramer1908 0-0 1 00 days Fig. 41. ditions there are too far removed from those of embryonic development to make it worth while considering this aspect of the subject in detail. The formula proposed by Faure-Fremiet for the growth SECT. 2] AND WEIGHT 399 of the epithelium of the foetal lung is, however, of greater interest i W = At^w + Bfiw + Ctw + Dw, where w is the total weight of the lung at the time in question. But here the number of arbitrary constants is so large that we reach the point where the question naturally arises whether an empirical /q^ 0,1 p 7 8 3 10 i1 12 13 1'^■ 15 16 17 18 19 20 21 Fig. 42. formula is worth looking for at all. The more complicated it is, the less valuable it is, in view of the fact that, in any case, it is not intended to give us an idea of the basic factors underlying the process. 2-5. Percentage Growth-rate and the Mitotic Index We have so far been examining the results of those investigators who have taken the curves obtained by simply plotting the weight of the embryo each day during development against the time, and who have endeavoured to find a correct mathematical expression for them without a preconceived theory. We have now to consider 400 ON INCREASE IN SIZE [pt. iii the work of those who have infused more theory into their treatment of the experimental facts. Before 1890 there was no regularity in the way in which experimentalists examined their data on growth. But about that time Minot began a long series of investigations on the growth of animals, mainly the guinea-pig and the rat, in which he introduced a new method, namely, the evaluation of the growth-rate by taking it as the increment in per cent, of the weight of the animal at the beginning of the period in question. Some workers, e.g. Preyer, had already adopted this plan. The percentage growth-rate has always been found to decline enormously as development proceeds, an observation which led Minot to say that the embryo gets oldest most quickly when it is youngest. This apparently paradoxical statement drew a good deal of attention to his work at the time, and his book. The Problem of Age, Growth, and Death, included many such graphs showing how rate of senescence was greatest in the earliest periods. One of these is reproduced here as Fig. 43. Another contribution of Minot's was the conception of the "mitotic index", or the number of mitosing cells per 1000 cells in a tissue. He did not himself find time to do more than a few of these laborious counts, but he gave the following figures, which showed that the mitotic index declined with age: Development of rabbit foetus (days) Tissue Mitotic index 7-5 Ectoderm 18 Mesoderm 17 Endoderm 18 lO-O Ectoderm 14 Mesoderm 13 Endoderm 15 Blood 10 130 Spinal cord Connective tissue II 10 Liver II Skin 10 Excretory tissue Muscle 6 6 These data lent weight to his principal conclusion, which was that the younger the embryo the more rapidly it aged. Practically nothing more was done along these lines until Olivo & Slavich in 1929 reported a large series of figures for the mitotic index of the developing heart in the chick. SECT. 2] AND WEIGHT 401 Davs Mitotic period in calculatec development index hours minutes 38 minu 2 22-5 19 42 — 3 21-2 20 55 — 5 169 26 1 1 15-7 7 150 29 24 94 9 8-3 53 II IO-2 II lO-O 44 3 4-8 13 7-0 70 15 57 15 4-4 100 33 40 17 3-7 117 39 2-5 19 2-8 155 55 4-2 21 1-2 363 7 1-8 10 (after hatching) 00 — The fall in the mitotic index ran closely parallel with the fall in the percentage growth-rate of the organ, as determined by a special series of weighings. The time taken by one mitosis was calculated from these data by Olivo & Slavich : it turned out to be constant at 38 minutes. But the intermitotic period grew longer and longer, indicating that the later growth consists less than the former of proliferation and more of increase in size of the cells already formed. o o\ ( Males ] Z 5 811 17 23 29 3S3« 45 75 80 105 120 135 150 165 180 185 210 dstye 241 Fig- 43 For a long time Minot's way of looking at embryonic growth in particular and growth in general was universally adopted, e.g. by Jenkinson, and even at the present time it is much used. But the Minot curve is undoubtedly based on a fallacy, and it was not long before a feeling that this was so began to arise. It was perhaps intensified by the appearance of such estimates as that of Muhlmann, who worked out the growth-rates in early stages of N E I . 26 402 ON INCREASE IN SIZE [pt. iii embryonic development in man as 3650 per cent, and above. It was not unnatural to enquire whether the Minot growth-rate of the original dividing egg-cell was even finite. The dissatisfaction was voiced in 191 7 by d'Arcy Thompson, who wrote as follows: "It was apparently from a feeling that the velocity of growth ought in some way to be equated with the mass of the growing structure that Minot introduced a curious and (as it seems to me) an unhappy method of representing growth in the form of what he called 'percentage-curves'. Now when we plot actual length against time we have a perfectly definite thing. When we differentiate this LjT we have dLjdT which is of course velocity, and from this by a second differentiation we obtain d^LjdT^, that is to say, the acceleration. But when you take percentages of jv, you are determining dyly and when you plot this against dx you have —-^ or dx — ^ or - . ^ , that is to say, you are multiplying the thing you wish y.dx y dx to represent by another quantity which is itself continually varying, and the result is that you are dealing with something very much less easily grasped by the mind than the original factors. Minot is of course dealing with a perfectly legitimate function of x and y and his method is practically tantamount to plotting logy against x^ that is to say, the logarithm of the increment against the time. [Cf. log. weight-age curves.] This could only be defended and justified if it led to some simple result, lOr instance if it gave us a straight line, or some other simpler curve than our usual curves of growth". This criticism was justified, for the Minot curve is certainly no simpler than the untouched growth curves ; it merely falls instead of rising. But d'Arcy Thompson did not point out the presence of a definite fallacy in Minot's way of looking at growth, a physiological rather than a mathematical one. This was grasped by Brody, who has written as follows: "Minot's method for computing growthrates gives an exaggerated decline in the percentage rates of growth with increasing age simply because an arbitrary unit of time, e.g. a week, does not have the same significance at different ages. It is, for example, entirely appropriate to express the gain in weight during a week as a percentage of the weight at the beginning of the week (Minot's method) for a 6-month old chicken, because the weights (i.e. the number of cells or other reproducing units) at the beginning SECT. 2] AND WEIGHT 403 and end of the week are nearly the same as compared to the gain in weight. But to express the gain in weight during a week as a percentage of the weight of the body at the beginning of the week for a 7-day old chick embryo would be quite fallacious. It would correspond to expressing the gain in population in the U.S.A. from 1666 to 1927 as a percentage of the size of the population in 1666. The growth of the population of the U.S.A. in 1927 is proportional to the population in 1927 and not to the population in 1666. Similarly the number of cells produced in a 7-day old chick embryo should be functionally related to the number of reproducing cells (i.e. the weight) of the chick at 7 days of age and not to the number of cells at I day of age. In brief, growth is a continuous process and the rate of growth at every instant should be functionally related to the number of reproducing units at the given instant and not to the number of reproducing units which existed in some relatively remote past". In other words, Brody would prefer to ask not how much 100 gm. of embryo add on to themselves during the immediately succeeding period, but rather how many grams of those 1 00 gm. had been added on during a short preceding period. Murray's modification of Minot's method, in which the mid-increments instead of the daily increments are used as the basis for calculation, goes some way to meet Brody's criticism, for it enquires how much 100 gm. of embryo add on to themselves during half the preceding and half the following period, thus speaking in terms of a more instantaneous measure. Brody himself has made use of a similar amelioration. However, Brody's real point is that the fault lies in choosing an arbitrary length of time interval for all stages of development, in spite of the fact that they cannot possibly be equivalent for the embryo. Brody might say that the embryo cannot be regarded as having been given an equal chance to accomplish its growth in each of the daily periods throughout its development. On the other hand, it might be argued that, though this is doubtless true as regards growth in weight, it is not true with respect to the activities of the embryo as a whole, which include many other processes, such as chemical differentiation. Taking the embryo as a whole in all its activities, the arbitrarily chosen and invariant period might be regarded as an adequate one. As we shall see later, this is essentially the same argument as that used by Murray against Robertson. 26-2 404 ON INCREASE IN SIZE [PT. Ill It may, however, be concluded that the Minot curve is only useful provided no theoretical conclusions are drawn from it, and that it is retained simply as a convenient method of comparing processes. Brody's own theories will be discussed later. As against the theory which Minot built up from his experiments with growing animals, Murray has brought forward one convincing -0.15 1 -0.14 n 4 «. I \ -0.12 '0.11 -0.10 -0.09 \ \ \ \ \ -0.08 ^ -0.07 -0X?6 • -0.05 \; > \ \ -0.04 0.03 ■0.02 K X X ^^ ^^--^ ^ '0.01 A^e 5 6 7 8 9 10 tl 12 13 14 15 16 17 IS 19 ^ Fk ?-44 argument. Minot's theory of cytomorphosis involved the following propositions: (i) that the rate of growth depends on the degree of senescence, (2) that senescence is at its maximum when development begins, (3) that the rate of senescence decreases with age, and (4) that death results from the differentiation of cells. But, as Murray says, we have no real evidence to show us that the "degree of aliveness" at any given moment is in any way connected with the velocity of growth at that particular moment, or, more correctly, that the latter SECT. 2] AND WEIGHT 405 value is a true measure of the former, "There are other and perhaps more significant phenomena", said Murray, "than the growth rate, which change with age." Murray himself proposed the use of a variant of the usual Minot curve by differentiating twice instead of once so as to get the acceleration and not the velocity. Thus, after having found the percentage growth-rate by the equation dt[_w\ t' where K ^ 3-6, he went on to find the negative acceleration for each day during embryonic growth : dw d dt dt = - Kt^. This value, so obtained, is, as it were, the negative increment of the percentage growth-rate, and shows a regularly declining curve (for the chick) as in Fig. 44. Such a curve must obviously suffer from the same disadvantages as the curve from which it is derived, and does not escape from Brody's criticism that the arbitrarily chosen time-units are incommensurable at different developmental stages. In 1922 Przibram observed that in many cases of post-embryonic growth the curve obtained by calculating according to Minot's method was extremely like a regular hyperbola, but he did not find that this was true for any example of embryonic growth. We shall see later what further use of this idea has been made, 2-6. Yolk Absorption-rate Another way of regarding embryonic growth (of a lecithic Ggg) is to concentrate attention on the whole system, instead of upon the growth of the embryo alone. It was in this way that Gray treated the development of the trout embryo in the paper already mentioned (p. 369 and Appendix i). He assumed that the rate of growth of the embryo was proportional to the dry weight of the embryo and to the dry weight of the remaining yolk. This idea had already been introduced for the trout by Kronfeld & Scheminzki (see p. 369 and Fig. 41) but Gray's figures were much more complete, and showed very clearly a falling off of growth towards the end of 4o6 ON INCREASE IN SIZE [pt. iii pre-natal life, when the yolk was becoming exhausted. Thus the wet weight of an embryo plotted against the time gave an S-shaped curve, which, however, was not symmetrical, for it had a point of inflection after about 70 per cent, instead of 50 per cent, of the development had been completed. This was of course well shown on the increment curve, which was skewly bell-shaped. Assuming that growth was proportional to the amount of yolk remaining as well as to the size of the embryo already formed, Gray developed an equation ^ (where x is the weight of the embryo at time t,yQ the total yolk in the unfertilised egg, K^, the amount of yolk combusted by one gram of embryo divided by the constant k in the equation dx J It = '"^' X and y being weight of embryo and weight of yolk respectively) which he considered accounted very well for the observed facts. He deduced from it that there should be a period at the end of development when the wet weight of the larva (the whole system, embryo plus yolk) is decreasing, although the wet weight of the embryo itself is still increasing. During the major part of development the wet weight of both would increase, owing to the absorption of water from outside. From the equations the maximum weight of the larva should be reached when o-86 gm. of yolk is still unconsumed, and in actual fact Gray found the peak at a point when i*io gm. \^as yet remaining. Another possibility used by Gray to test his hypothesis was that as it was unlikely that the temperature coefficient of the growth process would be the same as that of the catabolism going on, there ought to be a measurable difference in the size of fishes raised at various temperatures at the end of their development. Experiments designed to reveal such differences gave the following results : Mean weight of 100 embryos at Temperature (°C.) the end of incubation (gm.) 15 i3-35±o-i6 ID i507±o-i8 so that the higher temperature not only accelerated the process, but, by accelerating the combustion more than the storage, led to a SECT. 2] AND WEIGHT 407 smaller finished fish. These results are curious, for it is generallyunderstood that temperature changes alter the rate at which a growth-process goes on yet not the amount of end-product formed. The work of Barthelemy & Bonnet on the frog is an exact parallel to that of Gray on the trout, for these workers raised frog embryos at different temperatures with the following results : (P.E.G.) Temperature Dry wt. of 300 Dry wt. of 300 Dry wt. of embryos (°C.) eggs (gm.) embryos (gm.) Dry wt. of eggs 8 0-378 0-334 0-88 10 0-824 0-423 0-51 14 0-594 0-318 0-54 21 0-708 0-346 0-49 Temperature Dry wt. of 70 Dry wt. of 70 Dry wt. of embryos (°C.) eggs (gm.) embryos (gm.) Dry wt. of eggs 8 0-092 0-079 0-85 10 0-128 o-ioi 0-79 14 0-097 0-082 0-84 21 0-107 o-o88 0-83 (P.E.C.) If the second and third columns of this table are compared it will be seen that in the first series the French workers did get results like those of Gray, i.e. the higher the temperature the greater the combustions and the less the storage, but that in the second series there was no such effect to be observed. The Plastic Efficiency Coefficient (P.E.C.) is the most convenient way of expressing this relation (see Section 6- 1 o) . According to Gray it should change with temperature, for assuming his trout embryos to have the same percentage composition, no matter what the temperature, those raised at 15° would contain (each) 21*3 mgm. solid and consequently (since the eggs contain 43-4 mgm. solid) would have a P.E.G. of 0-50, while those raised at 10° would contain (each) 24-2 mgm. solid and consequently would have a P.E.C. of 0*56. I shall return to this subject in the section on general metabolism of the embryo; here it is only necessary to remark that the subject is clearly not yet settled and requires much more attention than it has so far received. At the same time, returning to the main theme, it must be remembered that Gray only made use of these temperature phenomena as one of the supports for his theory. He drew another support from the fact that if his equations were correct, the product obtained by multiplying the dry weight of yolk by the dry weight of embryo should be at a maximum on the 71st day of development. This he found to be actually the case, as is 4o8 ON INCREASE IN SIZE [PT. Ill shown in Fig. 45. And the growth-rate per gram of embryo (Minot growth-rate divided by 100?) was also proportional to the amount of yolk remaining. Gray himself indicated a number of criticisms which might be brought against his views. Thus there is little a priori reason for supposing that the growth-rate of the embryo should be determined by the amount of yolk in the yolk-sac, for apart from anything else, the 10 20 30 40 50 60 70 Fig. 45. Days of development. 80 90 100 amount and nature of the syncytium in the yolk-sac wall might be a limiting factor. One would also expect that in the beginning when the yolk is very large compared to the embryo, the yolk would be present in excess, and would not exert any influence on growth-rate. 2-7. The Autocatakinetic Formulae The workers mentioned so far might be divided into three groups, those who have elaborated empirical formulae, those who have adopted the Minot point of view, and those who have treated the SECT. 2] AND WEIGHT 409 yolk plus embryo as one growth-system. Now a fourth and very large group consists of those who have been greatly impressed with the similarity which some empirical growth-curves show with the curve for a monomolecular autocatalytic reaction. This manner of looking at the subject is associated mainly with the name of Brailsford Robertson, who has in many papers and in a book specially devoted to the matter put it forward as the most fundamental approach to growth. Robertson was not, however, the first to notice the likeness. As early as 1899 it had been referred to by Errera, and Ostwald a little later. According to Monnier, Chodat of Geneva paid some attention to it in 1904. "One may regard growth, as M. Chodat has suggested", said Monnier, "as a complicated chemical reaction in which the living cell is the catalyst and the substances present are water, salts, and CO2." Four years later (in 1908 on May 9) Ostwald's monograph on growth was published in the form of an inaugural dissertation, and only ten days later Robertson's first paper appeared in the Archivf. Entwicklungsmechanik. Ostwald had treated the question in a rather unmathematical manner, but had fully explained the nature of his hypothesis ; Robertson, on the other hand, gave the S-shaped curve a detailed mathematical treatment. " The carrying out of a progressive development in time has in animals a single characteristic type; the rapidity of the process begins at a low value, increases with the continuance of the action and falls off again at the end, in other words the type of curve is S-shaped." This was as far as Ostwald went, but he did not fail to point out that the S-shaped curve was identical with that of an autocatalytic or an autocatakinetic reaction. An important point to note is that Ostwald's curves were all curves of absolute weight — he did not in any instance plot increment curves. Robertson began by an explanation of the mathematical properties of the autocatalytic curve. The differential equation characteristic of the initial stages of an autocatalytic monomolecular reaction is as follows : , -r = k-iX [a — x), which expresses in mathematical symbols the fact that the velocity of the transformation is, at any instant, proportional to the amount of material which is undergoing change and to the amount of material which has already undergone change. If, however, the reaction has 4IO ON INCREASE IN SIZE [PT, III proceeded so far that the depressant effect of the products is measurable, then the previous equation becomes — = k^x [a — x) — k^x^. Now when the reaction has proceeded half-way to equilibrium, i.e. when X = la, the equation becomes X or log log A - X A — X = Ak{t- tj), Amount transformed and this is the well-known equation for the S-shaped curve, where X represents the body-weight (not the increment) at time t, A represents the maximum of final weight which the organism is to reach, ^1 the time at which half this maximum body-weight has been attained, and K a constant which has to be determined from a known value of a; at a given time t. By differentiating, it may be found that dx/dt is at a maximum when X = \A — in other words, that the rate of increase in weight is greatest when half the autocatalytic curve has been passed through. Robertson proceeded to apply his calculation to the growth of white rats, figures for which had been reported by Donaldson, Dunn, & Watson, with excellent results over part of ^^^' 4^* the curve, though not when the age amounted to more than lOO days. At the time it seemed as if this was a most convincing example of the value of the autocatalytic theory, but maturer consideration showed that this lOO days was but a third of the possible life-span, and the comparison of the two curves as made by Lotka, for instance, does not look so impressive. SECT. 2] AND WEIGHT 411 Fig. 46, taken from Robertson's first paper, illustrates the relation between the curves for the three chemical reactions, while in Fig. 47 is seen the theoretical and the experimental curve compared. In this same first paper, Robertson applied his autocatalytic equation to the growth of man, frog, a vine, and to certain organs. The frog figures, which were those of Davenport, were the nearest approach 50 150 200 Age in days Fig. 47 300 to embryonic growth dealt with by Robertson, and they are given in Table 54. They showed a good measure of agreement, but whether it was right to say, as Robertson did in his summary, that "in all probability cell growth or the synthesis of cytoplasm is an autocatalytic reaction" is a question that subsequent workers have not by any means answered with a bald affirmative, Robertson pointed out that his new interpretation of growth curves fitted in very well with the views that had already been advanced by Loeb. Loeb had suggested that the processes of cell-division and 412 ON INCREASE IN SIZE [PT. Ill growth were simply expressions of a more or less rapid return to the chemical equilibrium between nucleus and cytoplasm which had been temporarily shifted through the process of fertilisation. He thought that there occurred in the sea-urchin's egg during the early stages of its development a great synthesis of nuclear material, and that this synthesis progressed, roughly speaking, all the more rapidly the more nuclein was formed. The velocity of nuclear synthesis, said Loeb, increases with lapse of time in geometrical progression. Further, various observers, using the Q^k, formula, had concluded that the growth-process had a "chemical temperature coefficient", so Robertson felt fully justified in speaking of a "master-reaction" of growth, and in thinking of it as autocatalytic — a reaction, which, because it would be slower than any other, would act as the limiting factor of growth, and would impress its own particular character on the general appearance of the whole process from outside. Table 54. Larval growth of frog [tadpole). Robertson's figures Davenport's calculated ft-om Days after experimental autocatalytic hatching figures (mgm.) formula (mgm.) I 1-83 1-64 2 3 4 2-00 2-03 — — 5 3-43 3-90 7 8 5-05 6-00 9 10-40 9-OI 14 23-52 23-46 41 lOI-OO 110-90 81 1989-90 112-00*

"Obviously another growth-cycle supervening", said Robertson In his later writings, Robertson added further demonstrations of the applicability of the autocatalysis equation to growth, but he also added a large amount of extremely speculative considerations, such as the "nutrient level", the "endogenous catalyser", etc., into which we cannot here enter. Some of his suggestions, for example that the autocatalyst is lecithin, may be regarded as now definitely out of court. He showed, however, that the pre-natal growth in man (using Zangemeister's figures) was susceptible of description in an autocatalytic curve with its early " autokinetic " phase (Robertson's terminology), and its late "autostatic" phase. He also used, for the SECT. 2] AND WEIGHT 413 last part of the curve, data on the weight of new-born infants, born at different times, some earHer and some later than the normal. We have, moreover, already seen that the curve constructed by Vignes for human embryonic growth shows an S-shaped conformation. In 1926 Robertson published a long paper in which he reviewed the work which he and his collaborators had done on the subject of the autocatalytic theory of growth. Here he showed that the fall in the relative value of the velocity constant of (asymmetric) autocatalysis during the embryonic growth of the mouse followed almost exactly the same curve as the fall in the chemical nucleoplasmatic ratio as determined by LeBreton & Schaeffer (see Section 10-2). It is not clear, however, what significance is to be attached to this finding, especially as Crozier and Brown have shown that at least two velocity constants must be postulated in a given cycle. Robertson himself drew no theoretical conclusion from it. Robertson regards the several growth-cycles distinguishable in the life of an animal as being independent, in that they each have a different catalyst. The first of these in the case of the mouse has an equation of the type where x is the growth attained at time /, A the maximum growth attainable in the cycle under consideration, t the time required to attain half the maximum growth and k a specific velocity constant. B is another specific constant, an index of the asymmetry of the curve. The second and third cycles have the formula log-^-^ = A(^-0, where the symbols are as before. He considers that skewness or asymmetry originates probably in a progressive diminution of the velocity constant as described above. Robertson's suggestions were not allowed to go uncriticised. Meyer, in his report of data for the growth of the human foetus, took occasion to attack both them and the traditional Minot standpoint. On the whole, Robertson's autocatalysis theory emerged with less damage than Minot's cytomorphosis theory. Meyer brought into the light what had been one of the most disturbing features of the Minot method, namely, the absurd values which the percentage growth 414 ON INCREASE IN SIZE [pt. iii rate has in the earliest stages, e.g. five hundred and forty bilHon per cent. But, as Meyer seems to have completely failed to understand that Robertson and Minot were using different methods in calculating the "growth-rate", his criticism of Robertson was not of any importance. On the other hand, he did draw attention to the fact that growth-curves can be very misleading if it is not remembered that, though in the early stages the absolute growth is minute, the relative growth is enormous. "The weight of the impregnated human ovum", said Meyer, "is approximately 0-005 ^§"^-5 ^.nd yet investigators in all seriousness indicate its weight on a short ordinate reading in grams or even hundreds of grams. Little wonder, then, perhaps, that Robertson, Ostwald, and Read have unwittingly assumed that the curve of growth in man and mammals hugs the abscissa for several months as the curve of autocatalysis does." This was a good tilt at a common fallacy, but Meyer did not point out that it could be remedied by using log. paper, and he left it quite open to Robertson to reply that, even when strictly comparable quantities were taken, the S-shaped curve or a succession of S-shaped curves still resulted. More serious criticisms than these have, however, been brought against the Robertson method of treating embryonic growth. Luyet has pointed out that it may suffer more than the other methods from illusory difTerences in material. Again, Murray has written as follows: "(i) The formula demands the introduction of three separate constants which must be separately determined for every set of figures collected. (2) The equation does not give the weight as a function of age throughout life but only during an arbitrarily selected part of the growth-cycle. For these two reasons the equation as a practical simplification is not of great value. If the equation were in such a form that knowing the species, the age, the T°, and other environmental variables, one might calculate the weight and growth-rate, it might be of use. But as it stands now, it is necessary in each case to collect complete statistics and then find a mathematical expression of the figures obtained. For instance, all three constants in the equation for the growth of South Australian males differ from the constants used in the equation for the South Australian females. As one cannot extrapolate, the formula, like the man with one talent, returns what it receives. In fact as it covers only a section of the growth-curve, it yields less information than the SECT. 2] AND WEIGHT 415 original data. Moreover, as a rational account this description of the synthetic processes of growth is misleading, since (3) by this theory the growth-rate is proportional to the increment gain in weight regardless of the weight of the organism, or in other words, disregarding the amount or concentration of the reacting substances. (4) Figures for the growth of colonies as well as of individual organisms are said to be described by autocatalytic equations and are classed together. When growth is expressed in terms of percentage increase in mass, however, the important distinction between phylogenetic and ontogenetic growth is made evident ; for the former, after a short latent period, in the presence of an experimentally modified environment proceeds at a constant rate [cf Richards], while the latter does not. The S-shaped curve is the result of a limited and unrefreshed culture medium. The individual organism, however, instead of maintaining a constant growth-rate shows from the beginning considerable negative acceleration. (5) The S-shaped curve is not specific, for there are some physico-chemical processes not considered to be autocatalytic which are described by a similar curve. Finally, and this is the main objection, (6) chemical diflferentiation is not taken into account by the autocatalytic theory, which is based on the conception that there is some one master monomolecular reaction, which, being the slowest of the chain of reactions concerned in the phenomenon of growth, determines the velocity of the entire process. As there is no direct way of measuring the product of the masterreaction, the increase in the body-weight of the whole embryo is taken to represent the product. In view of the marked changes in chemical constitution which take place in the tissues with age there is no reason to suppose, and in fact it is extremely unlikely, that the total weight can be taken as an index of the amount or concentration of any one chemical substance". The tendency that has existed in the past to take the simple weight of the embryo as the sign par excellence of its developmental stage and its "aliveness" is only another case of the reluctance to judge by "ensemble", as Broca called it, instead of by single indices, which was so long the bane of physical anthropology. In this case, it is the increasing application of chemical methods to the embryo which has shown the superficiality of regarding mere weight as the pre-eminent factor. Lotka has drawn attention to this point of view, and is inclined to compare embryonic growth to the growth of a population such as Pearl & Parker have studied 4i6 ON INCREASE IN SIZE [pt. iii in Drosophila. Janisch treats the S-shaped growth-curve as the reciprocal of a catenary exponential curve. These presentations have the advantage that they do not prejudge the issue from a physicochemical angle. Exception to Robertson's views has been taken on quite other grounds by Snell, who points out that Robertson's equation ^ = K^Ax - K^x" dt ^ ^ and Crozier's modification of it ^ = (r^ + K,x) {A-x) (where x is the concentration of the end-product at time t, A the concentration of the substrate at time t, and K^ and K^ the velocity constants of the forward and reverse reactions respectively) do not take into account the fact that the system concerned is not a closed one. All the time the embryo is growing it is also eating, i.e. absorbing nutritive material, and in addition it is giving out waste products. Accordingly these equations derived from the law of mass action as we know it in the inorganic world do not allow for the effect of increasing size on the concentration of the reagents involved in growth. The equations hold true only on the condition that the volume occupied by the resulting substances remains constant, and since a growing organism is constantly increasing in -volume, this condition is not met. "When a chemical process is carried out in the laboratory," said Snell, "the reagents are ordinarily dissolved in water or some similar solvent, and the volume of the solvent is kept constant throughout the whole process. To make the conditions of a laboratory process comparable to those involved in the synthesis of new protoplasm the volume of the solvent would have to be increased as fast as the amount of the end-products is increased. As the solids of new protoplasm are formed, they do not stay in the same little parcel of liquid occupied by the old, rather they cause the liquids to expand with them. Hence the volume occupied by the end-products of growth is proportional to their amount, and the concentration of these products, instead of increasing, remains constant. This is a very important difference, for it is on the concentration and not the amount of reagents that reaction velocity depends." Thus the equations of Robertson and SECT. 2] AND WEIGHT 417 Crozier would only be true if the chick embryo, for instance, began as a kind of watery ghost of dimensions equivalent to those it normally has at hatching, and if development consisted in the gradual accumulation of solid substances within it. This singular kind of preformation certainly does not exist in reality. Snell developed another equation, ^ (where A is the initial amount of substrate, x the amount of endproduct at time /, ex the corresponding concentration at time t, Vq the volume of the organism at the beginning of the growthcycle, and iTi and K^ the velocity constants as before), which he regards as the correct form for the representation of an autocatalysed monomolecular reaction in which the conditions are similar to those in a growing animal, or in other words, where the volume occupied by the reagents increases in proportion as the end-product increases. Snell then showed that the curves obtained by this equation did not resemble any of the empirical growthcurves in the literature, and therefore concluded that there is no sound basis for assuming that the master reaction is either monomolecular or autocatalytic. Thus in the case of embryonic growth, where X — is almost constant, the curve approximates to that of a non autocatalysed monomolecular reaction, and again to nothing that is given by any actual embryo. In any case the only instance where a sigmoid curve has been shown to fit the growth of an individual foetal organ is in the work of FaureFremiet & Dragoiu on the lung of the embryo sheep. Robertson's point of view was adopted in the earlier work of Brody and his collaborators, who, however, introduced new viewpoints into it. Instead of plotting the absolute weights of the embryo against time, they plotted the increments per time-unit, thus obtaining curves similar to those in Fig. 34 above. Obviously, in a case where the absolute growth-curve was S-shaped, these increment curves would be doubly S-shaped, rising to a maximum at half-time and thereafter falling away. Or, put mathematically, from the differential equation I =Kx{A- x) NEI 27 4i8 ON INCREASE IN SIZE AND WEIGHT [pt. iii it follows that the velocity of change in an autocatakinetic system progressively rises from zero hour to a maximum value when x = ^A and afterwards constantly falls. When the data are plotted in this way the existence of growth-cycles, whether real or not, comes out much more clearly than when the absolute curves alone are used. It is often said, however, that the increment curve emphasises small fortuitous variations more than the absolute curve, and certainly there are cases, notably the chick embryo itself, where the increment curve shows up cycles which are but poorly shown on the absolute weight curve. Brody & Ragsdale in their first memoir on this subject dealt only with the growth of the cow, and concluded that one complete growth-cycle was accomplished in the foetal condition. A later paper considered the growth of the fowl, for which, on the data of Card & Kirkpatrick for growth after hatching, two cycles appeared, with maxima at 9 and 18 weeks respectively. For the growth of the embryo, the data of Hasselbalch and of Lamson & Edmond, re-arranged by Brody, gave two maxima also, but at slightly different times, thus: Lamson & Edmond 11-5 and 16-5 days Hasselbalch 10-5 and 15-0 days. Brody's figures for these curves are shown in Fig. 48. LeBreton & Schaeffer,who subsequently published a series of chick embryo weights, found maxima at 9 and 15 days, but in Murray's series there is no peak at any time, except a doubtful one on the i6th day, and a plateau between the 1 2th and 15th days. Schmalhausen's data again, when the increments are calculated, show in the case of both series peaks at 10 and 12*5 days, with an additional one, in the case of his 1927 series, at 17 days. In the face of this consensus of evidence, it is not altogether easy to conclude with Murray that "Brody's rhythmic growth curves were due to chance variations". It is true that Lamson & Edmond; Hasselbalch; and LeBreton & Schaeffer used too few embryos in their work, but even in Murray's own work, where about 650 embryos were used, there is an unexplained drop between the 17th and 1 8th days, as well as a plateau between the 12th and i6th days, both quite outside the probable error. Murray did not, it is true, get the i8th-day drop in all his experiments. In Schmalhausen's two series about 400 embryos were used. Brody's chick (daily increment curves) 4 5 6 7 8 9 ion 1213141516171819 20 3 4 5 6 7 8 9 10111213141516171S t t t t Lamson 8c Edmonds data Hasselbalch («) {b) Le Breton &. Schaeffer's chick (daily increment curve) ^ 765432i 4 5 6 7 8 9 10 1112 13141516 17 18 1920 W.Legh. | | (c) Schmalhausen's chick (daily increment curves) 7 N i^ ^ 5 6 7 8 9 10111213141516171819 4 5 6 7 8 9 10111213141516171819 1926 Max.t t 1927 t t t (d) (e) Fig. 48. 27-2 420 ON INCREASE IN SIZE [pt. m It may further be argued that weighing is a very simple process, and it is, therefore, difficult to see why errors should arise which should reflect themselves in these rhythmic curves. A greater degree of scepticism would be justified if they were the results of a complicated estimation method for a chemical substance. But, as it is, these curves form perhaps the best evidence which at present exists for the applicability of the Ostwald-Robertson view to embryonic growth. In the absence of a really exhaustive statistical investigation of the growth of the chick in the egg, these rhythmic curves must be accepted for what they are worth. They would be more convincing if all the workers had found peaks in the same places, but the variation which exists is no argument against the reality of the phenomenon in view of the fact that different breeds of hen were used. Further work is greatly needed to clear up this question. If the peaks on the daily increment curve do turn out to be real, it may be possible to relate them to the peaks of normal mortality which Payne and others have studied, and which will receive further consideration later. (See Fig. 443, Section 18-2.) The autocatalytic curve has also been found by Robertson to fit the data of Stotsenberg already referred to for the growth of albino rat embryos, and a peaked curve is obtained when the daily increments are plotted against time. But, as will be seen, Brody's exponential formula also fits these data, and it is probably right to conclude, as McDowell and his collaborators do, that the figures are not sufficiently good to allow us to distinguish between the two formulae. They cannot be regarded as supporting, therefore, any particular theory of embryonic growth. 2-8. Instantaneous Percentage Growth-rate Brody introduced still another way of representing the facts. He defined the "genetic growth constants" of animals as being the same within each genetically identical group of animals, and as corresponding to specific velocity constants and equilibrium constants in chemical actions in vitro. It may be seen from Fig. 49 taken from Brody's paper that the mature weight of the animal in its life-span. A, is approached by successively decreasing gains in weight after the point of inflection of the sigmoid curve has been passed. The velocity of growth, therefore, declines in a geometrical progression with age. The normal animal reaches, as Brody puts it. SECT. 2] AND WEIGHT 421 under a given set of favourable conditions (much more constant, of course, in egg or uterus than outside), a mature weight which is characteristic of its own species, just as the product of a chemical reaction in vitro reaches under a given set of conditions a definite equilibrium concentration characteristic of its kind. The mature weight A was determined by Brody for a large range of animals Fig. 49 by a graphical method. Now, in this process of geometrical progression in which the increments in unit time are becoming progressively smaller, it is found that in each unit of time the gain made in percentage of the gain made in the previous unit of time is a constant. Thus, in the autostatic growth-phase of the rabbit, for example, the gain is, during each month, 78 per cent, of what it was during the previous month. Brody calculated out this constant, k (simple growth persistency), for a great many animals. It corresponds to the specific velocity constant in chemical equations. 422 ON INCREASE IN SIZE [PT. Ill Finally, B is the difference between the mature weight of the animal and the weight the animal would have had at conception (a minus quantity) if the whole of growth was representable by the curve for the autostatic or self-inhibiting phase. This genetic growth-constant also was found for many animals by Brody. From Fig. 50 it can be seen that, the higher the value of k, the more rapidly the mature value is approached (pigeon > mouse > rat > guinea-pig > sheep > pig > cow > man), and that the fact can be equally well accounted for on the assumption that a substance Yrs.cS^ Age (from blrbh) man 7 8 9 10 11 12 Mos.a' 60 64 68 72 76 12 16 20 24 28 32 36 40 44 48 52 66 Age (from conception) of animals OoJ Fig. 50. is used up during growth, or on the assumption — perhaps more likely in view of the work of Carrel and his collaborators on tissue culture (reviewed by Pearl) — that during growth a growth-retarding substance is produced according to the law of monomolecular change. The paper of Brody, Sparrow & Kibler was concerned with age equivalence. They showed that, with the aid of the formula previously established by Brody, W=^ A- Be-""^ (where W is the weight at age t, A a. genetic growth-constant, the mature weight, B another genetic growth-constant which increases in value with increase in length of the processes preceding the point SECT. 2] AND WEIGHT 423 of inflection, and k a third genetic growth-constant, the fractional decHne in the velocity of growth), it was possible to plot growthcurves for all animals to the same base, and so to determine their age-equivalence. Thus they found that i rat month was equivalent to ii-gi cow months, and i guinea-pig gram to 509-1 cow grams. They finally constructed a table (Fig. 51), in which the value of ^ was given for a great many animals, and a logarithmic graph, from 001 .02 Value of k. .03 .M .05 06 -07 .C8 -09.1 Z 5 .6 7 .8 .9 10 CotVC. Mos; Fig. 51 which can be read off the time in months (conception age) at which the animal with the constant k in question will arrive at 10, 20, 30 or 90 per cent, of its mature weight. Brody next considered the growth-constants during the autokinetic or self-accelerating phase of growth. He subjected the methods which had previously been used to represent growth to severe criticism, part of which has already been referred to. Thus, in the case of Minot's method, increments of growth are regarded as being added on discontinuously at the terminal points of arbitrary time periods, whereas growth is really a continuous process, and I dt_ \ w =-k 424 ON INCREASE IN SIZE AND WEIGHT [pt. iii the mathematical expression for it must take account of its smooth nature. Brody found the relationship between the relative rates of growth k ^j^ W to be k = log [R + i), where R is Minot's percentage growth-rate. The latter can, therefore, only be used when it does not exceed 10 per cent, for the period under consideration, i.e. never in embryological work. Brody also criticised the methods of Pearl, whose equation dW _ k dt t- a does not take account of the fact that part of growth is self-accelerating, and also Pearl & Reed's modification of the original Robertson equation. He maintained that the best way was to plot the log. weight against the age, when, if the result is a straight line, the rate of instantaneous growth, A;, must be a constant. This will not of course be confused with the fractional persistency constant referred to above, for the latter only refers to the autostatic or self-inhibitory phase. The autokinetic or self-acceleratory phase is ^ clearly the more important ^ and interesting for embryologists. Figs. 52 a, b, taken from Brody's paper, show the data of various workers for the wet weight of chick embryos treated in the way described, namely, the log. weight plotted against the age. It will be noted that a series of straight lines result, forming a system concave to the abscissa and rising rapidly but more quickly at first than later. The curve is thus the exact opposite of the Minot Cms ,.:?

«jy 6 10 12 14 16 Ifl 20 Incubation Age Fig. 52 a. Gms. 20 10 </^ o I -H vy I I ^ ^ /. \ 6 5 3 — = = E —^ ? V % h — — — — — 1^ — / 1 z t — — . A /. / ?i // ^ ( J // [■ / / V — — — ^ .a ~ 1 — i^ — 1 — / K/ i — — — — — — ' — — y. / j 5 'if / / / "l ^^ / J \ .3 .2 ! ^^ A / i ^ — ^ ^ — — z — 1 / / — r i— — — — — — — — / J r^' H as SQl Da ch 1 /A B— rtr: r* Days 2 4 6 6 10 12 14 16 Incubation Age IS 20 Gms., 22 20 19 16 14 12 10 8 6 4 2 10 12 14 16 15 20 Incubation Ag<^ Fig. 52 b. 426 ON INCREASE IN SIZE [pt. iii curve, which falls so markedly during the same time. Thus the growth-rate expressed in this manner also falls off as time goes on, or rather rises less and less rapidly, becoming eventually asymptotic to the mature value. This curve is undoubtedly a great improvement on Minot's, for it involves no arbitrary time period and depends on the differential calculus, which has as its special province the evaluation of instantaneous change. An infinitesimally small period dt, and an infinitesimally small increase of weight dW, are the basis of its operation. Then all the infinitesimal differences can be added together (i.e. integrated). Thus: .j^ "^ _ becomes, if the number of dfs, and dW's is infinitely large [n) , when integrated, W = Ae^*, for I + ^y = e^\ A being the weight at the beginning of the whole period. If this is turned into logarithms log W =\og A ^ kt, , _\ogW - log A or k , and, as where growth is being considered A is, to all intents and purposes, o, ^ ^ k = -^ — . t The instantaneous relative growth-rate for a unit time is the sum of all the instantaneous rates during the given unit of time, and may therefore be multiplied or divided, according to the time-unit in which it is desired to express it. The log. weight/age graph is, therefore, a measure of the instantaneous growth-rate, and the value of the constant k which can be calculated from the last equation will give the slope of the straight, or approximately straight, line. The log. weight/age graph could, of course, have been plotted by Minot, but Brody's use of the differential calculus was required to show that the slope of the curve gave an instantaneous growth-constant. Thus, in the example given, Lamson & Edmond's data, the constant is 56 from the 5th to the 8th day of SECT. 2] AND WEIGHT 427 development (very steep slope), 36 from the 8th to the 13th day (less steep slope), 24 from the 13th to the i8th day (still less steep), and 25 from then onwards. The higher numerically the constant k, the steeper the slope, and consequently the greater the instantaneous growth-rate. In Figs. 53 a, b and Table 55 are shown most of the weights and processes in the hen's egg whose constants have been calculated by Brody. Each system grows at a rate peculiar to itself. Murray, as we have seen, also plotted log. weight against age, but he did not get a straight-line relationship ; on the contrary, the resulting curve was concave to the age (abscissa) . McDowell, again, got a similar concave curve for the pre-natal growth of the mouse, and there is much point in his criticism of Brody 's work: "Brody draws a series of straight lines through corresponding exponential curves and concludes that growth-rate does not decline continuously but by abrupt drops between periods of uniform rate. Since any curve can be approximated by a series of straight lines, the critical significance, both of the specific number of straight lines, and of his general conclusions, seems somewhat questionable"*. Table 55 includes also a column in which the time taken for the embryo or a corresponding entity to double its weight or amount is shown. For, when the instantaneous percentage growth-rate is constant, the time intervals between doubling of weights are constant; therefore, from the expression W - Ae^\ at a certain time logs _ 0-695 k ~ k ' and, as k is found to be for the rat embryo 0-53 or 53 per cent., the time required for it to double its weight must be — ^ or 1-3 days. ^ 0-53 Further, if growth in weight can be taken as a measure of the increase in the population of cells in the body, a new cell-generation is produced every 1-3 days on an average, and the cell-division frequency is 1/1-3, i-^- 0-77 times per day. It is thus possible to determine, as Brody says, the mean life of a mother cell before it divides into two daughter cells.

Nevertheless, McDowell himself admits a discontinuity between pre-axial and axial growth, as we have seen on pp. 394 and 396. n c i 55 Mom s.Gms 3.0 20 20 10 1.0 a .8 5 .5 3 .3 2 .2 1 .1 .001 .01 DaysO 6 10 12 14 16 18 20 Incubation A^ Fig. 53«' ^1 8 S88 S '^'^^ ^ ^ ':ind:^no'oO 430 ON INCREASE IN SIZE [PT. Ill Table 55. Instantaneous growth-rate {k). Brody's figures. Time in Growth which the Time rate % entity is Entity in question (days) per day {k) doubled {d) Investigator Chick Wet weight . 5-8 56 1-2 Lamson & Edmond ,, 8-13 36 1-9 JJ 5> 13-18 24 29 5> 55 6-10 56 1-2 Hasselbalch 10-14 29 2-4 55 14-19 19 3-6 35 6-10 47 1-5 Murray 10-14 33 2-1 35 14-19 21 3-1 33 CO2 excretion 0-4 98 07 Atwood & Weakley >> 4-14 31 2-2 35 35 5> 14-19 Pause — 33 33 3) 19-21 31 — 33 55 0-16 36 I '9 Hasselbalch 5) 0-16 32 2-2 Murray Urea excretion 5-7 76 — Needham ,, 7-14 34 — 53 Glutathione content 6-9 54 — Mvirray 5) 9-15 30 — 55 Total CO2 content 7-9 56 — 55 ,, 9-16 37 — • 35 Chloride content 12-15 32 — ,j Calcium content 12-16 85 — Plimmer & Lowndes Creatine content 14-21 28 — Mellanby Nitrogen content 6-9 60 — Murray jj 10-15 47 — ,, 15-20 23 — Total solid content 5-10 57 — ,, 10-16 46 — Ash content ... 10-14 43 — ,, 14-19 25 — Calorific value 7-9 56 — ,, 10-15 47 — Rat Wet weight ... 13-22 53 1-3 Stotsenberg Guinea-pig Wet weight ... 17-20 100 0-7 Ibsen & Ibsen; ,, 20-35 25 2-8 Draper; Hensen ,, 35-52 9 7-8 S3 55 ,, 52-70 5 15-1 33 33 Man Wet weight ... 60-110 8 8-7 Streeter ,, 110-160 Not straight 55 ,, 160-240 1-7 41-0 33 , 240-280 1-3 55-0 „ Brody himself did not omit to make suggestions as to possible correlations between his abrupt breaks in growth-rate and other phenomena known to be taking place during the embryonic development of the chick. In the first place, he associated the breaks in the growth SECT. 2] AND WEIGHT 431 rates of carbon dioxide production at the 14th day with the change in mode of respiration from aquatic to terrestrial which takes place late in incubation. This is quite a convincing correlation, but his suggestion that the first break (at four days), before which the instantaneous growth-rate is about 100 per cent., and the Minot growth-rate 1000 per cent., is associated with a general critical period occurring at that time is not really so satisfactory. For almost any process has its critical moments during development — for example, the peak in protein metabolism at 8-5 days. In cases where there is no a priori reason for assuming correlations except the one fact that their peaks coincide or are converse to each other, the utmost caution should be used in so correlating them. Wholesale correlations of apparently unrelated phenomena may be chemically misleading. Thus Brody cites Tomita's peak in total lactic acid content at the 5th day (see Fig. 292) as evidence of a critical period corresponding to the abrupt break in his growth-rates of carbon dioxide production and to the peak in Payne's mortality curve (see Fig. 443). Brody is not the only investigator who has occupied himself with the growth-rates of different chemical processes and amounts in the embryo, but, before passing on to discuss these points, which will lead naturally to the question of the growth-rates of parts of embryos, a further word must be said about Brody's work. At present it is not possible to tell much from the comparison of embryos of different kinds, though it is obvious that an immense field of research is opened up here for the comparative embryologist of the future. Thus the equation for the development of the chick embryo in weight according to Murray is W^=o-668^^^, corresponding to instantaneous growth-constants of 0-47, 0-33, and 0-2 1 successively,* while the equation for Stotsenberg's rat embryo figures, according to Brody, is W^ = 0-000065^°^^*, corresponding to a steady rate of 53 per cent, per day instantaneous. On the steadiness of this rate Brody says, "If there is no fallacy in this reasoning we have reached a new and an extremely important conclusion. While all investigators of the time relations of growth have reached the conclusion that the percentage growth-rate continuously and rapidly declines with age, our conclusion is that the instantaneous percentage growth-rate remains constant for the relatively enormously

A later value, due to Vladimirov & Danilina, is W=o-'^2^fi'^. 432 ON INCREASE IN SIZE [pt. iii long period between 14 days and birth. The cause of this difference in results is due to the fallacy in the method of analysis employed by Minot". This is only true, subject to confirmation of the fact that the foetal log. weight/age graph gives a straight line over definite periods, and this is just what is not certain. Decision on the matter cannot yet be made. Another interesting point which emerges from Table 55 is the long embryonic stage in the guinea-pig. The chick hatches when its k is about 0-21 and the rat is born when its k is even higher— 0-53, but the guinea-pig stays inside the uterus until its instantaneous percentage growth-rate has dropped to 0-05. Brody succeeded, indeed, in raising guinea-pigs by feeding them on hay and grain immediately after birth, so that they tasted neither colostrum nor milk. It is interesting, again, to note that there is only one break in the instantaneous growth-rate of carbon dioxide production, whereas there are at least two in the instantaneous growth-rate of wet weight. This must mean either that the respiratory function develops at a rate quite independent of the growth in mass, or that the weight of the body cannot be taken as an index of the growth of the metabolising tissues. This point will be referred to again, for it is of much importance in chemical embryology. On the other hand, the respiration k does show a break about the 17th or i8th day, which is duplicated in the wet weight k, or, at any rate, in the log. weight curve constructed from Lamson & Edmond's data — for it is not so apparent in those of Hasselbalch and of Murray. This may be associated, as we have seen already is Brody's suggestion, with the change in form of respiration occurring then (chorio-allantoic to pulmonary). There is no doubt that some obscure events are associated with this late stage in the chick, e.g. the mortality peak of Payne, which can be greatly intensified if a certain lethal gene is present, and the sudden immunity to implanted rat sarcomata (Murphy), which the chick then acquires. Brody suggests that the chick embryo passes at this stage through a "metamorphosis" similar to those hidden ones which exist, according to Davenport, in the development of man. The extremely small values of k for the embryonic period of man are worth attention. The human embryo grows a great deal more slowly than any other. Five months after conception the instantaneous percentage growth-rate is only 1-7 per day, while, during the week preceding birth, the rat embryo grows at the rate of 53 per SECT. 2] AND WEIGHT 433 ■ whole embryo (Murray and Needham) Bcalorlfic value (Murray) Ddry solid (Murray) Ocarbohydrate (Needham) ® protein (Murray and Needham) e fat (Murray) cent, per day. The lowest rate of growth ever reached by the rat after birth is 3 per cent, per day. Given percentage rates of growth, therefore, do not signify equivalent stages of development irrespective of the species of animal. Calculation of the rates of growth for various processes and individual components in the development of the embryo has also been done by other investigators using the Minot method. In 1927 I calculated the percentage growth-rate for the total carbohydrate content of the chick embryo; it fell from 56 to 22 per cent. In Fig. 54 is shown the fall in the Minot curves for the wet weight of the whole embryo, the calorific value, the dry weight of the whole embryo, the sugar, protein, and fat content of the embryo. All of ^° them fall, but we have here an instance of the limited but real value of the Minot curves, which, although no absolute conclusion can be drawn from them, do show that the tissue constituents and the dry weight have a different behaviour from the wet weight. It can easily be seen that they form a plateau between the loth and the 15th day, during which they grow at a constant rate while the wet weight is falling all the time. This plateau also appears on the curves for carbon dioxide output calculated in the same manner as percentage growthrates from the figures of Bohr & Hasselbalch; Atwood & Weakley; and Murray in 1927 by Brody. The plateau must be due to the fact that the growth of dry substance is specially rapid during the middle phase of development; it is then that the embryo makes the most rapid strides from wetness to dryness. It is interesting to see that the growth-rate of carbohydrate is never as high as some of the others, and never drops so low. It is significant, moreover, that on the 1 9th day the Minot growth-rate of the protein has dropped below that of the whole body, while the growth-rate of fat remains well above it. This is an illustration of the "relative" use of Minot's method. Fig- 54 88 434 ON INCREASE IN SIZE [PT. Ill 2-9. Growth Constants Brody is not the only worker who has applied the differential calculus to embryonic growth-curves. Teissier and Lambert & Teissier suggested simultaneously that this should be done, but their work was quite theoretical. However, Schmalhausen published independently at almost exactly the same time a paper in which it actually was done. He criticised Minot's method of calculating Per cent per ^"y k=.53 50 ^ k- ^ \nW2 IntVi 5 40 _ ^ V. ^ i. 0,30 — ct JO J3 (0 3 t. 0. 0> I JS20 C c 'i3 Q) w U JP Q. k=.11 c ID — c. k=-047 U=-031 ^ .^^ Rat,?, — 1 1 h unmated 1 1 >== l_ 1 1 1 1 1 1 1 1 1 Dayso 10 20 30 40 50 60 70 so 90 100 no 120 iso uo i50 16O i70 18O c o O 8 CO 18 28 38 58 68 78 Age Fig. 55 a 108 118. 128 138 U8 158 " mittlerer prozentualer Zuwachs " in arbitrary time-units from exactly the same point of view as Brody. Thus, he says of the Minot method: "the larger time intervals we take, the bigger the error will be. With equal time intervals, the error will be bigger the bigger the rapidity of growth, and this will in fact lead to altogether misleading figures for the early periods". We need not follow Schmalhausen's reasoning, which led him to adopt the calculus as a better assistance in studying growth-curves, for we have already examined and approved the SECT. 2] AND WEIGHT 435 C JO J3 a 10 ll f \ / \ ^ / / 1 \ \

A ^ c r c=— ,^ arguments of Brody. Schmalhausen speaks of the "wahre Wachstumsgeschwindigkeit " instead of the instantaneous percentage growth-rate, and of r instead of k. His equation relating the instantaneous percentage growth-rate to the Minot growth-rate is exactly the same as Brody's, and he points out that, when, according to the old method, the growth-rate would be 700, the instantaneous method would give a result of 207, though 50 per cent. (Minot) would be equivalent to 40-5 (instantaneous). These figures might have been read off from the graph of relation given by Brody, and it is surprising that the two workers, one at Kiev and the other in Missouri, should have been thinking on such very similar lines. It is still more surprising that embryologists had not thought on such lines long before. Schmalhausen and Brody diverge, however, upon one important point, namely, the shape of the line given when the log. weight is plotted against the age, for Schmalhausen regards it as a curve — just as Murray and McDowell do — while, as we have seen, Brody lays great stress on the representation of it by a series of straight lines having abrupt breaks between them. Thus, the instantaneous percentage growth-rate, which with Brody remains constant over certain definite periods, with Schmalhausen continually declines in value. In other words, Brody's diagram which shows the instantaneous percentage growth-rate dropping in a stepped formation from fertilisation to hatching is replaced in Schmalhausen's work by a regular curve passing downwards to become asymptotic to the abscissa (Fig. 55 h), just as the Minot curve does, only, of course, plotted from a set of figures having a real meaning as against the abstractions of Minot. This difference of outlook leads naturally to very wide differences in conclusions; thus Schmalhausen has nothing to say about critical points or hidden metamorphoses. Having diverged from Brody in this direction, 28-2 10 n n 16 20 Zl Age in days P=wet weight; Cv = instantaneous % growth-rate. Fig. 55 *• 436 ON INCREASE IN SIZE [PT. Ill he proceeded a good deal further along it by observing that the graph relating instantaneous percentage growth-rate to age was practically identical with a rectangular hyperbola, and that there was a simple relation between the values of r or Cv (Brody's K) and the age, for the product of the two was always roughly equal to 300.* Table 56. Embryonic growth: Schmalhauseri's '■^Wachstumskonstante" {^^wahre WachstumsgeschwindigkeW^ x time). Cvl Cut length age Investigator Man Whole embryo wet weight 193 369 Friedenthal Mouse 95 5> — 337-5 McDowell et al. Rat 35 J5 441 518 Stotsenberg Chick J> » — 318-5 Murray 35 55 518 321 Schmalhausen Liver 329 Lung ... — 321 Fore limb (whole period) — 293 (3-18 days) ... — 329 Hind limb (whole period) — 347 (3-18 days) ... — 395 Stomach ... — ■ 374 Brain — 210 Lens — 210 Whole eye (2-10 days) ... — 317 ,, (i 1-21 days) ... — 99 Heart — 276 Mesonephros (4-13 days) — 224-5 Metanephros (7-17 days) — 359 (17-21 days) — 196 Ovary — 145 Guinea-pig Whole embryo wet weight — 347 Draper Trout 55 55 (30-51 days) — 206 Kronfeld & ScJ (51-99 days) — 207 S3 Schmalhausen gives no explanation of the breaks in the cases of those organs which have two values of Cvt, but calls attention to the fact that the organs of early differentiation have low Cvt and vice versa. If the curve obtained by plotting Cv (Brody's k) against time is a regular hyperbola, then the product Cvt should be 300. If it exceeds this figure, the curve is descending and becoming asymptotic less rapidly, i.e. the rate of growth (instantaneous) is not falling off as rapidly as it will be if the product is less than 300 at any given moment. This constant he calls the "Wachstumskonstante", and its values, calculated by him for a number of embryonic processes, are seen in Table 56. It is perhaps the least convincing part of his exposition, for when during a certain series, e.g. the growth of the human embryo, the constant Cvt oscillates between 899 and 93 as extreme limits, one may legitimately doubt whether great stress can be laid

Brody himself does not find this to be so. SECT. 2] AND WEIGHT 437 on the average. Moreover, Janisch treats the same curve as a catenary exponential one. And, ahhough the " Wachstumskonstante " for the various organs and parts of the embryo show differences which might well be regarded as characteristic for the tissue in question, it is disturbing to find so wide a difference from the predicted value in the case of the rat embryo, explained though it is by Schmalhausen as due to variable factors in the food of the maternal organism. The reason why 300 is the number to which these figures approach is, of course, because, according to Schmalhausen's formula, the increase of the embryonic weight can be expressed by the equation W= k [atf, where W is the weight, a the "Lineargrosse", t the time and k a constant. This agrees with the hyperbolic nature of the Cvjt curve. Table 57. Instantaneous percentage growth-rate [Chick). Schmalhausen Day of Brody j^ development (Smoothed) (Raw) (Smoothed) O-I 1-2 2-3 — — — — 190 190 3-4 — 119 140 4-5 — 139 107 5-6 — 83 87 H 47 79 70 7-8 47 38 60 8-9 47 36 50 9-10 33 53 '^2 lO-II 33 22 38 11-12 33 ^\ 33 12-13 33 48 30 13-14 33 30 27 14-15 21 20 25 15-16 21 32 23 16-17 21 23 21 17-18 21 22 19 18-19 21 II 17 19-20 — ^l 15 20-21 — 16 12 for the equation of an equal-sided hyperbola is j; = 3/x. Schmalhausen does not derive his Cv directly, but calculates it in each case from the Minot percentage growth-rate figures. It is instructive to place side by side the instantaneous percentage growth-rates of Brody and Schmalhausen for the chick embryo, as is done in Table 57. That of the former has three constant periods, that of the latter 438 ON INCREASE IN SIZE [pt. m shows a gradual decline, and the figures illustrate what has already been said, namely, that, until we possess much better statistical data than is actually the case, we cannot differentiate between the Brody position and the Murray-McDowell-Schmalhausen position.* As the matter is fundamental in view of the important theoretical issues involved, the accumulation of more data is urgently to be desired. It may be mentioned that Cohn & Murray, plotting log, weight/age curves for the growth of embryonic heart cells in tissue culture, obtained curves concave to the age axis and not straight lines. Schmalhausen also studied the growth in length of the chick embryo, calculating it from the weight by the formula L = VW. The daily gain in length a he found to be variable around a constant value of 1-47 for the first half of development and 2-00 for the second half. But when the weights were corrected by the estimation of the embryo's specific gravity (average for first half 1-025, average for second half i -06) the corresponding daily gains in length worked out at I -Go and 1-79. A further correction made necessary by the presence of the feathers during the last half of development brought the figure down to 1-64, so that throughout incubation the embryo apparently grows in length at the same average rate. The duck embryo, according to Schmalhausen, has a daily length increment of i*io mm., and this value he regards as constant for the species. He went on to calculate a for the human embryo, using the weight data of Friedenthal and Zangemeister, and for the embryo of the white rat, using the data of Stotsenberg. All these results are shown in Table 58, together with his further assessments of a calculated from the " Normal tafeln" of Minot & Taylor for the rabbit and Keibel for the pig. The daily size ("Lineargrosse") increments of his own measurements of separate organs and parts of the chick embryo are also given. During the course of development the value of a rises and falls according to the rate of growth. If the value a/ 1 is calculated, where a is the daily increment in length and / the length of the part in question at the beginning of the period, the absolute size of the part will cease to affect the result, and the organs will be comparable with themselves and with the whole embryo. When this is done, the ratio is found to be fairly constant, rising as high as 10-3 per cent, for the stomach and falling as low as 6-3 per cent, for the lens. These figures are also given in Table 58. Schmalhausen concluded from

Recent work by Byerly supports that of Brody. SECT. 2] AND WEIGHT 439 them that organs which reach a high state of differentiation early grow the most slowly (brain and lens), while less differentiated organs grow most quickly (liver and limb-buds). The growth of the body as a whole is the average practically exactly of the rest, and it is interesting to note that the organ which most nearly approaches it is the heart. The heart would seem to grow in size at the same rate as the entire body. But, as Schmalhausen says, this growth in size seems to have no simple relation to the growth in weight as shown by the percentage growth-rate. Table 58. SchmalhauserC s values for a, i.e. daily increment in size or '^ Lineargrdsse^\ Chick Duck Man )> Rat Rabbit Pig Guinea-pig Whole embryo (ist half) (2nd half) Brain Lens Spleen Heart Lung Liver Testis Metanephros Stomach Fore limb Hind limb Pectoral muscles Whole embryo Millimetres A r a a/l Investigator 1-6 Schmalhausen 1-64 7-65 05 6-50 o-i 6-30 o-i 6-62 03 7-50 0-29 7-82 0-55 883 0-09 8-82 0-32 ID- 10 0-64 10-30 0-44 7-33 0-76 885 0-43 909 I-IO — 0"55 — Friedenthal 0-55 — Zangemeister 1-47 — Stotsenberg 1-2 — Minot & Taylor i-i8 — Keibel 0-75 — Read In a later paper Schmalhausen studied the relation between initial weight and end weight in a number of animals, wishing to obtain some means of comparing their " Wachstumsertrage " or Growthyields, on a basis independent of their size. He found that u, or the mass of substance added on to itself by the organism between times ti and tz, could be calculated by the formula u = t h *^i 0-4343 u u post u embryonic embryonic whole period period life-span 2-2 17-2 19-4 1-8 15-3 I7-I 9-5 3-8 13-3 8-9 2-3 II-2 13-6 4-3 17-9 13-6 2-3 15-9 15-6 4-8 20-5 20-9 3-3 24-3 440 ON INCREASE IN SIZE [pt. iii the initial weight being taken as unity {k = Cvt). His results for various organisms show interesting differences, thus : Sturgeon Pike Hen Mouse Rat Guinea-pig ... Pig Man Here we observe the effect of early hatching in the two aquatic forms, which have the greater part of their growth still before them at the time of leaving the tgg. The other figures demonstrate quantitatively what is apparent to common sense, namely, that the embryonic period is the time of greatest growth in terrestrial animals. 2-10. The Growth of Parts We must now turn to the relative growth-rates of parts of the embryonic organism. This is a field which has mainly been tilled by anatomists, but it is of the greatest importance to the chemical embryologist. For the increasing and decreasing intensities of physicochemical processes cannot be intelligently studied in the absence of a knowledge of the distribution of the whole mass among the different organs and tissues. The investigation of the relative growths of endocrine glands, again, cannot but throw much light on the development of the adult metabolism in the embryo. D'Arcy Thompson sees an appreciation of this in the eighteenthcentury preformationists. "It was the apparently unlimited extent", he says, "to which, in the development of the chick embryo, inequalities of growth could and did produce changes of form and changes of anatomical structure that led Haller to surmise that the process was actually without limits and that all development was but an unfolding, an 'evolutio' in which no part came into being which had not essentially existed before. In short the celebrated doctrine of preformation implied on the one hand a clear recognition of what, throughout the later stages of development, growth can do, by hastening the increase in size in one part, hindering that of another, changing their relative magnitudes and positions, and altering their forms ; while on the other hand, it betrayed a failure — SECT. 2] AND WEIGHT 441 inevitable in those days — to recognise the essential difference between these movements of masses and the molecular processes which precede and accompany them and which are characteristic of another order of magnitude." The papers of Schmalhausen are of much importance in this matter. Inspired by the views of His, who declared in 1874 that all the development of shape could be ascribed to unequal growth in various component parts of the embryo, he set himself to weigh and measure a great number of these individual sections. He first studied the relative growth-rates of the brain and eye of the chick embryo, together with the liver, lung and stomach, representing the organs of endodermal origin. In each case, he calculated the % growth-rates and the percentages formed of the weight of the whole body. For the organs of mesodermal origin, he chose the heart, the mesonephros, the metanephros, the ovary, and testis. These figures he treated in the same way. In many cases his weights were not obtained directly but by reconstructing from serial sections and then weighing, proper allowance being made for complicating factors such as specific gravity. Fig. 56 shows one of his graphs — it is specially interesting as showing the definite decrease in weight which the mesonephros undergoes after the 15th day in giving place to the metanephros or adult kidney. It also includes % growth-rate curves for the fore and hind limbs. Lastly, he ascertained the growth-rate of the feathers. In general, he found that the changes in the growth-rates of organs were synchronous. The percentage growth-rate (see Fig, 57) seemed to have peaks in its descent, each one less marked than the preceding one. In each case, the growth-rate of every organ shows a certain rise, but the amount of the rise differs in different cases — thus the lung is the organ which is growing fastest about the 6th day, the hind extremity about the loth day and the stomach about the 13th. On the whole, the periods of depression of the growth-rate of the majority of organs are from 7 to 9 days, from io| to ii| days, and from 14 to 16 days. When the weights of individual organs, however, were arranged plotted against weight of embryo, not age, the peaks disappeared, as would be expected, for the total weight is the sum of the weights of the organs. It would be interesting to plot the logs, of Schmalhausen's organ-weights against age in order to obtain the instantaneous growth-constants of Brody for each one. Schmal 442 ON INCREASE IN SIZE [PT. Ill hausen's general results, however, were as follows: on the 5th and 6th days, the growth-rate of all organs is falling, with the possible exception of liver and hind limb. This continues till the beginning of the 7th day, save that the eye and the lens may show a slight rise. At the beginning of the 7th day, however, the growth-rates of all organs rise, firstly the mesonephros, the liver and the lung, and, to a ^17 S. 16 x5 10 1.4 o^l3 o\i2 <n clO ^ 9 c m 8 2 7 (0 5 X3 E (E) = Fore limb G = Gonad M = Mesonephros Abs. web wei ghb 12 3 4 5 6? 9 10 11 12 13 14 15 16 17 18 19 20 Days Fig. 56. less extent, the hind limbs, then these are followed by the rest. After a peak all fall until the 1 1 th day, when all again rise, only to fall on the 14th, with the exception of the feathers, which maintain a rise. Later a more gradual rise in growth-rate takes place throughout the body, begun by the lens and liver and, to a less degree, by the eye and the brain, and continued by the stomach and the mesonephros. After the subsequent fall, only small variations take place, which are found to be synchronous for groups of organs such as kidney-liver SECT. 2] AND WEIGHT 443 stomach until the end of the embryonic period. It is certainly interesting that organs so different in origin and nature as the eye and the mesonephros should be similarly affected by spurts of growth at various stages, and Schmalhausen concluded that this was an argument against the hypothesis of specific organ-stimulating substances, the presence of which would from time to time cause more 1 Z 3 V 5 6 7 a 3 10'^ 1Z ^3 1'^ -75 16 17 13 13 20 Zi Days Fig. 57. C = brain; £; = hind limb; A" = whole body; Z-t = liver; Li = lens; Af=mesonephros; Oc = eye; P = lung; Af /n = metanephros ; 5^ = stomach; G= gonads. rapid growth in one place of the embryo than in another. It looks much more as if growth-promoting substances were passed round in the embryonic circulation at certain definite intervals, and so exercised an effect on a large number of different organs. In this connection, the recent work on the growth-promoting substances of egg-yolk should be borne in mind, and the experimentally determined cycles of varying permeability to fat-soluble and watersoluble substances in the walls of the vitelline blood-vessels. One relation which seems clear from Schmalhausen's work is that growth 444 ON INCREASE IN SIZE [pt. iii of fore and hind limbs does not accomplish itself in the same spurts as the viscera do, for during the 8th day depression in the growthrates of the latter the skeleton is growing vigorously, and during the loth day peak it rather falls off. Very similar remarks apply to the hind limb growth-rate. Schmalhausen concluded that very young organs can respond to a given intra-embryonic environment by increase in growth-intensity, while more differentiated organs can at the same time respond by depressions in their growth-rates. "If one and the same influence", he says, "can act in a stimulatory manner on the growth of some parts or organs, and inhibitorily on the growth of others, we can see how unequal growth can take place and hence a change in form." It would also appear that the more development goes on, the more different the rates of growth of different organs are. Three factors seem to control the growth-rate of a single organ: (i) the age of the embryo, (2) its own degree of differentiation, and (3) growth-promoting substances or embryonic hormones present in the circulation. Under (2) would be included the time of origin of its "anlage" and the intensity with which its preliminary growth would take place. These views are not compatible with Mehnert's "laws of organogenesis", the main one of which was that the growthrate of an organ in the embryo was proportional to its degree of development at the time in question. The only criticism that can be levelled against Schmalhausen's work is that the number of embryos employed was perhaps rather few. In conjunction with Stepanova, Schmalhausen made further investigations on the growth of the embryonic skeleton in the chick. Similar fluctuations in pre-natal growth-rates of parts have been discussed as regards the primates by Schultz. Schmalhausen has attempted to give an explanation of these spurts in terms of metabolism. Summarised again there are, in the case of the chick, three or four periods, in each of which the growthrate first rises and then falls, as follows: Days 0—4 1st period, great fall from a high value 4-9 2nd period, rising to the 6th day then falling 9-12 3rd period, rising to the loth day then falling 12-15 4th period, rising to the 13th day then falling 15—21 5th period, rising to the 17th day then falling He has suggested that these periods may partly correspond to the periods which can be distinguished in the development of the chick SECT. 2] AND WEIGHT 445 embryo, during which one type of chemical molecule is predominantly burned to furnish energy for the growing organism. This subject will be handled fully later (Sections 6-8, 7-7, and 9-5) ; here it suffices to say that the beginning of development is in many ways closely associated with an important carbohydrate metabolism, and the latter part with the metabolism of fatty acids, while the intermediate part would appear to have an association with catabolism of protein, in view of the fact that the point of maximum protein catabolism occurs when 8*5 days of development have been completed. These periods, in Schmalhausen's view, may be identified with those in which he finds spurts in the growth-rate. A certain amount of scepticism about this identification would seem justifiable until we have irrefragable proof that the spurts are more than chance variations in a curve composed of too few data. His suggestions involve the view that "Abbauprodukte" accumulate from time to time in the developing embryo, and so hinder its growth (essentially the same theory as those of Jickeli and of Montgomery) . Thus his first depression of the growth on the 4th day corresponds to an accumulation of lactic acid and ammonia (see further on for the detailed references) and his second depression of growth on the gth day corresponds to an accumulation of urea. Finally, his third depression of growth about the 12th day corresponds to an accumulation of uric acid. He admits that there is nothing chemical which obviously coincides with the later depressions of growth, but supposes that they depend on the decreasing excretory power of the mesonephros. After the i6th day the metanephros would be undertaking the duty of excreting waste products, and growth accordingly begins again. Ingenious as these correlations are, they cannot be said to be convincing, in view of the fact that many other processes besides the excretion of waste products may be supposed to be exercising an eflfect on the growth-rate. More interesting is Schmalhausen's attribution of great importance to the surface of the blastoderm, the blastodermal capillaries, and the active surface of the excretory organs. Measurement of these during the course of development would throw a bright light on these problems. Schmalhausen did himself compare the growth in weight of the embryonic kidneys with the daily increment of the whole body, and, although the figures were rather erratic, he felt able to conclude that, owing to the slow growth of the mesonephros and metanephros, the excretory surface 446 ON INCREASE IN SIZE [pt. iii was only just keeping pace with the growth of the embryo. In these circumstances, it was not surprising to find now an accumulation and now a flushing out of waste products from the embryonic body. In a later paper, however, he modified considerably his views on this subject. As regards the growth of individual parts, Schmalhausen later introduced several further expressions. "Homonomic growth", in his terminology, means growth of an organism in which all the parts and organs have the same growth-constant, "heteronomic growth" — the more usual form — is the growth of an organism composed of organs each with its own characteristic growth-constant. Then the growth-directing force may exist either inside or outside the anlages of the separate organs — in the chick it apparently does not exist inside — and in the former case it would be called "autonomic growth", in the latter " automorphic ", while, if the influence was directly the growth of another organ, it would be termed "heteromorphic". Schmalhausen found that, although the organs in the chick embryo taken at any one moment had very different rates of growth, yet, if they were all dated, as it were, from the time of formation of their anlages, they showed very similar rates of growth. Thus an anlage developing late would be growing much quicker than the whole body, while, at the same time, if its instantaneous percentage growth-rate curve was plotted, it would be found to have a shape very like that of the organism as a whole. Thus organs can only be compared as to their growth-rates if they are taken from their own particular origins and not from the origin of the body as a whole. In homonomic growth, of course, one is dealing with organs originating at the same time and having identical growth-constants. In this case, the definite proportions of the resulting organism can be deduced from those of the anlages ; in other words, a kind of preformation holds good. If Table 56 be again referred to, it will be seen that, on the whole, it takes functioning organs longer to grow than functionless ones. Thus the metanephros, which at first has a Cvt of 359, drops to 196 after it has begun to excrete actively by about the i6th day. In Fig. 58 is shown the relation between the weights of the organs in the embryo chick expressed as percentages of the total weight of the body. The heart and mesonephros are seen to have their maximal relative size very early in development, after which the former declines slowly and the latter more rapidly. The first four days of SECT. 2] AND WEIGHT 447 development see also the maximal relative size of brain and lens, but these fall very rapidly away from their pre-eminence. Towards the end of the developmental period, the fore limb gains much in importance, and about that time also the metanephros reaches a maximal point of growth. For further comparison further calculations are necessary. The relative instantaneous percentage growth-rate could be obtained from the equation ^ , 1 ^ (^v _ log v-^ — log V Cw log Wi — log w ' 1 "\ \ / \ zv 22 20 18 IB n n 10 / \ \ \ I \ / \ s ' / A \ ,^ / \ \ / ^ •^ ^ 1 \ / \ \ 1 / \ s H 1 1 f ^\ ^A \ / 1 1 ""■- ^ ^ 8 6

I. /u ^ L^ ^ — — - — — ..^ ^ ^ — -^ ^

^ z £ ■ — xC — ' '^ =^ =; N ^--M "^ 1 I 3 V J 6 7 9 10 Tl 12 13 n IS 16 1? 18 IS ZO 21 Days Fig. 58. £ = fore limb; G = brain; ^=heart; Z,=lens; jV= metanephros; f/ = mesonephros. where Cv is the instantaneous percentage growth-rate for the organ or part in question, v-^ and v the weights of the organ at the beginning and end of the period in question, Cw the instantaneous percentage growth-rate of the organism as a whole and w-^ and w the weights of the organism as a whole at the beginning and end of the period in question. But this would not take into account the time of formation of the various anlages. More complicated expressions have, therefore, to be found, but as they do not at present seem to have any direct importance for the chemical embryologist, a reference to the original paper must suffice. They involve the computation of an "extensity factor" which is usually the same as the 448 ON INCREASE IN SIZE [pt. iii constant a, already referred to, and an intensity factor which is the corrected product of the instantaneous percentage growth-rate and the time, i.e. Cvt. The relative extensity factor is obtained by determining the time which is taken by the organism or the organ to grow i mm. in length, thus reversing the process by which a was originally found. The size of the anlage is also included, and called the mass factor. By the aid of all this apparatus, Schmalhausen compares organs on a common basis, i.e. the time taken for i mm. increase in length to be made. Thus the extensity factor of the chick embryo brain [r) is 1-27 and that of the duck embryo brain 1-26, which means that the sizes of the respective organs are in their earliest stages almost identical. But the relative instantaneous percentage growth-rate (intensity factor, k, or Cv) differs considerably, for in the chick it is 1-87 and in the duck 2-01, which means that the duck embryo brain grows distinctly more rapidly than that of the chick embryo and finally attains a larger size. Again, for the stomach of the chick embryo the extensity factor, r, is 0-244 ^^^ ^^^ the duck 0*324, but the intensity factor is 3-59 for the chick and 2-86 for the duck, or, in unquantitative terms, the stomach is rather bigger to start with (relatively) in the duck than in the hen, but the chick stomach grows faster and reaches eventually a larger proportion of the body. This work on disproportionate or heterogonic growth led Schmalhausen into a field which had been in course of investigation by Huxley and others. Schmalhausen was able to obtain Huxley's formula from his own, and concluded with some justice that his own were of fairly general validity and did not hold only for embryonic growth. On the other hand, owing to the absence of a true extensity factor in Huxley's formula, the latter could not be applied to the embryo; for, although in post-embryonic growth-curves the ages of all the organs can be taken as approximately identical with the age of the animal, this is by no means the case in embryological work, where the time of formation of the various " anlages " is of the greatest importance. The investigations on heterogonic growth are not immediately germane to the theme of this book, but they may at any moment become very important for the chemical physiology of the embryo, and it is necessary therefore to be aware of them. Schmalhausen's work is really an extension to the embryo of the conceptions of Pezard and Champy, as worked out in recent years by Huxley SECT. 2] AND WEIGHT 449 and his associates. A rich harvest awaits the investigator who discovers the relation between chemical constitution and the differential growth-ratios. Perhaps a fruitful line of work will develop from the finding of Robb that the log. weight of an organ plotted against the log. body-weight often gives a straight line. He has suggested that organ-growth may depend on a kind of partition-coefficient, organs competing, as it were, for the building-stones in the blood-stream, and securing now a greater now a lesser proportion, according to the changing permeability of their cell-walls. The changes which occur in the chick's relative growth-rates of parts at hatching have been studied by Latimer, who combined together the data collected for pre-natal stages by Schmalhausen and those for post-natal stages by various American workers. His results lead to the conclusion that the organs and parts fall into three groups : (i) Those in which no change in relative growth-rate is found, e.g. liver, gizzard, feathers, ovaries. (ii) Those which show a brief post-natal retardation, e.g. total body-weight, brain and heart. (iii) Those which show a marked post-natal acceleration, e.g. kidneys and spleen. As will be shown below, the brain and eyes, so prominent in the embryo, fall consistently throughout life when expressed as per cent, of the whole weight, while the gizzard, liver, kidneys, spleen and heart have a maximum in early post-natal life. Other work on relative sizes of parts has been done by Jenkinson on embryonic trout, by Keene & Hewer on man, and by Jackson who gives a graph (Fig. 59) showing the relative proportions in the human embryo at different stages of its development, collected from all the available data. Boyd (on man) and Welcker & Brandt (on the chick, salamander and man) made earlier attempts at the same thing, but the ages of their embryos were unknown. The graph demonstrates the relatively large size of the brain in the early sizes, and in many ways resembles the graph for the organs of the chick given by Schmalhausen. "In general," says Jackson, "the period of maximum relative growth passes in a somewhat wave-like manner over the body from the head towards the foot. The head reaches its maximum relative size about the 2nd month. In the trunk, the upper portion, including the thorax and the upper abdominal viscera, is relatively largest throughout the earlier half of foetal life. NEi 29 450 ON INCREASE IN SIZE [PT. Ill The lower part of the abdomen becomes more prominent towards the end of the foetal period, due chiefly to the rapid expansion of the intestines at this time. The pelvis and lower extremities do not reach their greatest relative size until early adult life, although the upper extremities have reached their maximum relative size at birth. It may also be noted that the organs lying dorsal to the body axis grow at first far more rapidly than those ventral to the body axis, for, 20 40 60 80 100 120 140 160 180 200 220 240 260 280 Days Fig- 59 while in the 2nd month the former are three times the size of the latter, at birth they are equal, and in the adult the latter are six times the size of the former." Jackson's data on the growth of such organs as the suprarenal gland should also be of much service to chemical embryologists, and his paper as a whole is of great value, as it summarises the results of all the earlier workers — Welcker & Brandt; Brandt; Anderson; Lomer; Meeh; Liman; Thoma; Oppenheimer; Collin, Lucien & Beneke; Devergie; Schmidtt and Elsasser. The general results of all the workers who have occupied themselves with the weights of foetal parts are tabulated in Table 59, which gives the point in development at which the maximum percentage of the total body-weight is reached, what that percentage is, what it becomes at birth and what eventually it is in the adult animal. SECT. 2] AND WEIGHT 451 Table 59. Relative maximal weights of constituent parts of the embryo. Point at which m ^ v I m 1 1 m 1 <i % of total body-weight Part or LLL€XJ\.LLkL\jixL\. 13 attained in At maxi At birth or At adult organ development mum hatching stage Pig Total viscera 1 5 (length in 38 160 8 (Lowrey) Head 1 8 mm.) 30 22-0 6 Brain i8 9 4-0 0087 Spinal cord i8 1-87 0-33 004 Eyeball 86 1-15 04 o-oi Heart 15 4-64 10 0-32 Lungs 86 3-9 20 07 Liver 25 15-9 31 1-38 Kidneys 58 2-59 I-OI 025 Mesonephros 15 120 — Spleen 15 — 0-17 0-13 Pancreas 15 — o-i6 0-14 Thymus 15 — 0-37 — Thyroid 58 — 0026 0-004 Suprarenal 58 0-13 O-OIQ 0-005 Intest. and stomach 15 3-6 4-79 Man Head 2 (months) 45 26-0 (Jackson *) Trunk I 65 40-0 — Fore limb 10 10 lo-o* Hind limb 10 20 200 i 72-4 Brain 2 20 135 — Spinal cord I 5 0-15 — Heart I 5 0-7 — Liver 2-5 7-5 5-0 — Lungs Spleen 4 ID 33 0-4 2-0| 0-4 i 0-94 Thymus 10 0-3 0-3 — Thyroid ID 0-I2 0-I2 — . Kidneys 7 i-o 1-05 Suprarenal 3 0-45 024 0-90 Chick Heart 4 (days) 1-5 0-56 — (Schmalhausen) Mesonephros 4 o-6i 0014 — Metanephros 16 0-39 023 — Brain 4 300 2-6 — Lens 6 0-17 0025 — Fore limb 16 32 2-1 Stomach 20 3-59 3-56 — Dogfish Head 0-09 (wt. in 40-0 17-5 12 (Kearney t) Skin (birth) 75-0 gm.) 11-3 I I -3 7 Skeleton 85 100 8-6 9 Muscles (adult) 630 45-0 63 Total viscera, (a) c. o-i 19-0 I2-0 — two maxima (b) c. 350-0 14-3 — 9-8 Brain 01 15-0 1-6 0-9 Spinal cord 01 1-76 0-5 0-17 Eyeballs o-i 90 2-0 064 Heart O-I 4-0 0-15 0-20 Pancreas 350-0 0-14 006 0-14 Liver 20-0 7-0 4-8 59 Spleen 350-0 0-38 0098 0-2 Rectal gland 01 01 05 0032 — Mesonephros 1-8 4-8 i-i 0-38 Testes and ovaries 200-0 0-9 0-4 0-28 Stomach and intest. 350-0 5-5 2-9 4-0

Including all previous work on man. t "The various organs and parts in dogfish [Mustelus cams] embryos and adults show relative growths strikingly similar to that which has been observed among the higher vertebrates, including mammals and man". 29-2 452 ON INCREASE IN SIZE [pt. m From this it can be seen that the organs which reach their maximum relative weight early in development are the heart, spleen, pancreas, thymus, brain, spinal cord and head. The mesonephros also, of course, reaches its maximum fairly soon and declines more or less rapidly afterwards. The muscle masses, shown especially in the figures for fore and hind limbs, increase steadily in relative weight and reach their maximal relative size at or shortly before birth. The suprarenal gland, the stomach, the lungs, and the thyroid are variable in their point of maximum. But, as can be seen from the table, the data on these matters are not very numerous, most of the attention which has been given to the relative weights of parts and organs having gone into the study of post-natal life. W. Schultze has made an interesting investigation on the effect of hormones on the developing parts and organs in the tadpole. It is interesting that the only conclusions to which Jackson would commit himself were ( i ) that the embryo grows much faster in the earlier stages than in the later, and (2) that, at any rate as far as vertebrates were concerned, pre-natal growth is relatively much greater at the cephalic than at the caudal end. These points had already both been stated by Aristotle, and the whole advance lay in giving them a quantitative backing. Jackson did not consider that his figures supported the view of Preyer that those organs grow fastest in the embryo which will afterwards first come into functional operation. A great mass of such data has since Jackson's time been collected by Calkins & Scammon and many investigators working under their influence. We heed not do more than mention their work on the growth of the spinal axis in the human embryo, that of Scammon on the height-weight index, Nafiagas on anencephalic embryos, and Brody and Hammond on proportions in the cow, for these and many others only indirectly concern us. But it is important to note that for the human embryo Calkins & Scammon found that from 3 months onwards the growth in length, girth and diameter of the various external divisions of the body was directly proportional to the growth in total body-length. While each dimension has its own growth-rate with respect to the total body-length, this characteristic rate does not alter during the period under consideration. All these entities then may be expressed by the Calkins equation D =^ aL±b, SECT. 2] AND WEIGHT 453 where D is the dimension in question, L the total body-length, and a and b constants. The constant ^ is a measure of the amount of growth that has gone on prior to the period in question, and, as it is negative for the limbs but positive for all head and neck measurements, the conclusion is that while the latter have been growing extremely rapidly before the period began the former have not. This is as would be expected. The same conclusion emerges from the data of Corrado on the weights of head, trunk and extremities as analysed by Scammon. Many organs have been examined by the investigators of this school. The cerebellum, for instance, was found by Scammon & Dunn to increase in absolute volume and weight first slowly and then more rapidly during the foetal period; thus, its percentage growth-rate rose for the first six months of pregnancy, only to fall sharply afterwards. The pancreas, studied by Scammon, grows at a rate very like that of the whole embryo, but the relative weight of the organ with respect to total body-weight undergoes a reduction from 0-3 per cent, at the 4th month to o-i per cent, at birth. The uterus, on the other hand, passes through two definite phases in pre-natal life. Until 7 months the organ shows a lineal increase with respect to body-length which is comparable to that of most lineal body-dimensions, but after 7 months it grows much more rapidly. In early post-natal life, however, the organ goes through an involution stage which has long been known, actually decreasing in size by hypoplasia and hypotrophy until it reaches the level it would have attained had the early foetal growth-rate been continued. "This suggests", says Scammon, '^that the growth of the uterus in the latter foetal months consists of a substratum of typical foetal growth plus a secondary increment due to an extra stimulus furnished by a hormone of placental or possibly ovarian origin." Here is an excellent illustration of how an organ can act as an index registering obscure physico-chemical changes in the internal environment of the embryo. The other organ which undergoes a reduction in size following birth in man is the suprarenal gland, and it also has been investigated by Scammon. But, unlike the uterus, its growth when observed in the foetal period shows no increased intensity towards the time of birth, so that the involution which occurs afterwards by degeneration of the two inner cortical layers decreases its size far below what it 454 ON INCREASE IN SIZE [pt. iii would have reached had it gone on growing at the same velocity as before birth. This leads on to the general problems raised by the growth of the ductless glands in the embryo, problems of the greatest importance in view of the regulating influence which the foetal endocrines probably exercise. The pituitary gland, according to Covell, grows proportionately to the total body-weight in human pre-natal life, i.e. slowly till about the 5th month and more rapidly thereafter. The thyroid also shows no outstanding variations from the normal curve. The growth of the thymus, however, is characterised by high variability, while the pineal gland grows at nearly the same rate as the brain. The results of studies on the weight of these glands, therefore, do not reveal any striking correlations, and they must be supplemented by histological evidence. This will be presented in the section on hormones. Scammon; Scott; and Scammon & Kittleson have studied the growth of the intestinal tract and the stomach in the human embryo. While this work does not give us any help in evaluating the active absorptive surface during embryonic life, its main conclusions are of interest. Thus the growth of the gastro-intestinal tract follows the law of antero-posterior gradient or direction, for the more cranial portions grow relatively more rapidly in the early part of foetal life, while the successive caudal portions show smaller amounts of growth at the beginning and larger ones later. The number of crypts and glands in the stomach mucosa seems to increase per sq. mm, very regularly during the progress of foetal growth. Watkins' study of the growth of arteries is also interesting, for it shows that the vessels which supply the foetus only have a rate of growth similar to that of the body as a whole, i.e. slow for a short time at first, and then for a long time rapid, while the arteries which supply the placenta as well as the embryonic body have a long period of slow growth followed by a short period of rapid growth. These facts throw a certain light on the metabolic needs of the developing organism. Much valuable information is contained in the papers of Scammon and Armstrong on the foetal growth of the eye, and in that of Noback on the respiratory system, but it cannot be given here. Davenport has recently considered the growth-curves of man in the light of the work of Scammon and his associates, and Gunther has written on these subjects especially in relation to sex. SECT. 2] AND WEIGHT 455 2-1 1. Variability and Correlation Enough has now been said about the growth of the parts and organs of the whole considered in isolation, and we must consider the relation between the growth-rate and two other factors, namely, variability and correlation. The population of cells in the metazoal embryo may no doubt be compared with the populations of protozoa in cultures, but, whereas the functions of the whole in the latter case are very limited, those of the whole in the former case are highly complex. In other words, one may enquire to what extent there is variability between different embryos of exactly the same fertilisation age. Closely allied to this question is what is, to all intents and purposes, its converse, namely, at what point in development is the correlation coefficient greatest, i.e. at what point is the swing of variation among embryos away from the mean least? It is to be regretted that these enquiries have not been very deeply carried on in embryology, but there are some rather significant observations which need attention. So far only the mean values for weights and measures of embryos have been under consideration. But obviously no statistical study of these is complete without a consideration of the amount of variability among the individual cases from which the mean value is derived. The variability coefficient is defined as the standard deviation X 100, mean the standard deviation being a measure of the spread of points around the mean, i.e. a measure of the point upon the frequencycurve where the change takes place between concave to the mean and convex to it. Fig. 22 showing McDowell's points will explain the meaning of this. It has long been known that the variability coefficient decreases with age in man, and it is always stated that it follows the changing growth-rate quite closely, but some confusion has been caused in the past by a doubt as to what manner of representing the growth-rate is being referred to. The fact is, however, that the variability coefficient follows the simple increment curve. Thus, if for absolute growth a sigmoid curve holds good, the greatest daily or monthly increment will occur as we have seen at the middle of the period, and this peak will coincide with a peak in the variability coefficient. This was found to hold in 456 ON INCREASE IN SIZE [pt. m actual fact by Boas & Wissler, by Boas, and by Bowditch, who studied exhaustively the growth of Toronto school-children. The variability coefficient in that case followed exactly the curve of yearly increment and reached an exactly simultaneous maximum at the age of 15 years. The correlation coefficient behaved in the same way. Boas & Wissler explained their results by saying that correlations between measurements in one individual ought naturally to be greater during periods of rapid growth than at other times, because the variations in responsible factors will affect them all to an equal extent. Variability is what governs correlation, so that would also be expected to rise and fall in the same manner. But there is a proviso that must be made here, for Boas & Wissler's use of the term "growth-rate" is not the same as that of Brody, for instance. Boas & Wissler mean by the time of greatest growth-rate the time at which the largest increments are being added on to the organism in unit time, i.e. the half-way point in the curve of the autocatalytic equation. Brody means by the time of greatest growth-rate the time at which the organism is adding on to itself the largest relative increments, i.e. the earliest stages of embryonic life, when the slope of the log. weight/age curve is extremely steep, and the embryo doubles its weight in an exceedingly small lapse of time. The variability coefficient shows, therefore, that it is at any rate true to say that, during the phase when the largest absolute increments are being made, the widest variations from the mean tend to occur. This seems very reasonable, but there is also evidence which shows that, when the period of most rapid growth in Brody's sense is occurring, the variability coefficient is also large. Other examples are numerous. Buchem found a coefficient of variability of 0-4 for the early stages of the embryo cow and o-i later. Edwards found a variability coefficient of 0-1347 for unincubated chick blastoderms, but of 0-1087 for those incubated 24 hours. Jenkinson gives a graph exactly analogous to those of Boas & Wissler computed from Roberts' measurements of English artisans. Jenkinson also worked on the trout embryo (or rather the alevin, for his first point was 2-3 weeks after hatching, by which time the yolk was not completely absorbed). He found that there was a close general agreement between weekly increment and variability coefficient for the first 10 weeks after hatching, true not only for the growth and variability of the body-length but also for some of the parts such as SECT. 2] AND WEIGHT 457 the diameter of the eye, the length of the head, and the length of the caudal fin. Fig. 60, taken from Jenkinson, shows how, although the curves do not run absolutely parallel, they certainly rise and fall together. On the other hand, the correlation coefficients between several pairs of organs show that in many cases — total length and breadth of caudal fin, total length and length of anterior dorsal fin, total length and length of head, head length and eye diameter — there is a significant diminution in value during the time that there 70 a .^ 60 iB 50 40 C « 30 c y 20 o 10 < Tot \ \ al lar AX /ae. Total 1 lengt h At \ \ ^ ^^V, \ \ \ V V J ^ \

% >

\ X^^^ ■» m -, .3 1'3 2'3 3*3 4*3 5-3 6*3 7'3 8*3 9*3 10-3 Weeks after hatching Fig. 60. is a decrease in growth-rate. Thus, when the growth-rate is highest, the variations between individuals are greatest, but the correlation coefficients between various organs or parts in the same individual are also greatest. It is easy to see why this should be so, but Boas has given a mathematical proof of the relation between these coefficients and the growth-rate. Expressed differently, it could be said that the faster the growth-rate the more proportional the growth but the greater the variation as between individuals. Turning now to the early part of the embryonic period, the first complete investigation was that of Fischel, who studied the individual variations between duck embryos at the primitive streak stage. Von Baer 458 ON INCREASE IN SIZE [PT. Ill had already averred qualitatively as early as 1828 that variability was much more pronounced in the earlier stages than in the later ones. Kupffer & Benecke; Keibel & Abraham; and Assheton afterwards drew special attention to it. Fischel, however, measured the length of the embryos, and found that the older they were the more regularly they agreed together. This led him to conclude that regulating influences came into play during development which brought about a more synchronous course of growth and differentiation, and made 6 7 8 9 10 n •^ Number of Somites Fig. 61. the individual variations less and less obvious. Such a standpoint would accord well with frequency-curves such as that of McDowell and his collaborators for the mouse embryo (see Fig. 22), where the range of weights on a given day is 4 or 5 times as large at the beginning of development as it is at the end. His and Levi, working on the development of the chick, came across the same phenomenon. Fischel divided the total length of the embryo into a number of constituent lengths, e.g. from the cephalic to the caudal end of the somites and from the extreme cephalic end to the anterior blastopore ("Darmpforte"). He distinguished 13 such lengths, and these he measured in a large number of embryos, judging the age in each case by the number of somites formed. As will be seen from Fig. 61, the limits of variation considered as absolute maxima and minima SECT. 2] AND WEIGHT 459 are much the same when the embryo has o somites as when it has 20, but, in view of the increase in total length during this time, it will be seen that the maximum variation is much less important as time goes on. "The relative differences", said Fischel, "are certainly less in the later stages than in the earlier ones." Fischel's measurements of the lengths of the parts all showed the same relation as regards variability, but, though the length of the body is increasing regularly all through this period, the length of the part between the anterior end and the ist somite remains practically stationary, as does the length of the part between the last somite and the posterior end of the embryo. In other words, the increase in length is entirely due to growth of the middle region in which the somites are being produced.* The size of the individual variations can be large; thus an embryo may be more than 50 per cent, longer than another one of the same stage. Fischel's examination of the work of other authors, such as that of Bonnet on the sheep and of Keibel on the pig, induced him to suppose that very similar effects were seen in mammalian embryos. Philiptschenko carried the question into insect development by investigating an apterygote, Isotoma cinera, and found that the older stages showed the greater variability. On the other hand, Zuitin, who has studied the development of Dixippus morosus from this point of view, found that, just as in the birds and mammals, the earlier stages were the ones which showed most variations from the mean. Schmalhausen has also considered the question on the basis of the figures he obtained in his studies on the growth of the chick embryo, already referred to. He points out that the early stages in embryonic development are those when the anlages are being formed. At that time, every few hours, as it were, are marked by the start off of one or more parts or organs on the long declining hyperbola which represents their instantaneous growth-rate. Thus a cross-section through an embryo in those early periods would demonstrate, if some method of the future made it possible, a series of growth-rates, some low, appertaining to the more senior organs, some high, appertaining to the more junior ones. Moreover, the mass of each anlage is different for each organ, so that extremely complicated effects will be observed if the weights of the organs are described in percentage of the total body-weight. Thus

As Levi has shown, embryo size in the somite stage is very similar no matter what the size of the adult bird, but soon the larger animals grow longer. Also the size of the first somites of all sauropsida is between 5000 and 8000 /x^ surface. 460 ON INCREASE IN SIZE [PT. Ill 10-0 the earlier the stage the more chance there will be for individual differences in growth-rate to reveal themselves, while in later development these will be equalised, adjusted, and compensated by a process of self-regulation. Schmalhausen does not say, however, what this process of self-regulation is, and its nature, indeed, offers one of the most interesting problems in embryology. We shall refer to it again in the chapter on hormones. Schmalhausen suggests that Philiptschenko's results might be explained by the extremely small time elapsing between the formation of the various anlages in animals with ultra-rapid incubation times such as some insects. Owing to the high instantaneous growth-rates in the earliest stages, a minute difference in time of formation of an anlage or a minute difference in its initial size will make a very large difference between one individual and another. 2*12. Explantation and the Growth-promoting Factor We must now return to consider further the essential nature of the embryonic growth process. Some reference has already been made to Schmalhausen's findings with regard to spurts of growth and the extent to which all organs are affected simultaneously by the growth-impulse. But what is g ?• this growth-impulse? What is its .£ e-o essential nature? It is no longer ^ g.o possible to say that we have no J ^ idea, for there are a number of very significant observations regarding it arising from the study of tissue culture in vitro outside the living body. Ephrussi has brought forward evidence suggesting that growth-rate may be a genetic factor. The study of the effect of temperature on the growth-process has also been fruitful to a certain extent in answering this question, and its results will shortly be described. In the first place, it is natural to enquire whether the growth-rate found to hold for the entire embryo growing under its normal conditions would hold for parts of it isolated and growing in tissue 4 & 6 7 8 9 10 1112 13 14 15 16 17 18 Incababion age Fig. 62. SECT. 2] AND WEIGHT 461 culture. Would small pieces of it be found to grow in tissue culture proportionately to the growth of the organism as a whole, or is the growth-rate a value absolutely dependent on the intactness of the embryo? This question has found a definite answer in the remarkable work of Cohn & Murray, who, following up a few precursory experiments of Carrel, removed hearts from chick embryos from the 4th day onwards, made 1-0 Days 5 J I L l—J I 1 ¥ J dw db W cultures of them in adult hen plasma or plasma plus Ringer solution, and determined their rate of growth. This was done by four methods : {a) subtracting from the total area of tissue after n days of cultivation the area of the central portion, {b) dividing the area of growth made (total minus central portion) by the size of the original fragment, {c) dividing the area of growth made by the size of q the central portion at the time of measurement, and [d) dividing the area of growth made by the square root of the area of the central fragment. With these methods and carefully controlled conditions they found that the growth-rate of the cultures of embryo heart fell in an exactly parallel manner to the Minot curve previously found by Murray to hold for the chick. Figs. 62 and 63, taken from Cohn & Murray's paper, show this fall, and may be compared with Fig. 34 which shows the corresponding Minot curve for the whole embryo. The kink in the heart culture curve may perhaps be related to special changes occurring in the organ at the 9th day (see Section 23). It must be remembered that the heart of the chick has been stated by Schmalhausen to grow at the same rate as the whole body, though, as a reference to Table 56 shows, the Schmalhausen Cvt is 276 for the heart and 321 for the whole body. However, it would not have been material which organ had been selected by Cohn & Murray, for all organs have growth-rates either descending 6 7 8 9 10 11 12 13 14 15 16 Incobation age Fig. 63. 462 ON INCREASE IN SIZE [PT. Ill in steps (see Fig. 55 a) as Brody would have it, or in a hyperbola (see Fig. 55^) as Schmalhausen would have it. That Cohn & Murray compared their results with the Minot curve does not matter, as all percentage growth-rate curves, whether instantaneous or not, fall towards the abscissa asymptotically. Cohn & Murray's achievement was to demonstrate that the growth-promoting impulse resides to a large extent actually within the cells, and that its fading out can be seen equally well in the decreasing rapidity with which isolated cells will grow in culture as in the decreasing rapidity with which the whole cell-population of the embryo increases in the egg. In other words, functional changes taking place in the organism as a whole are mirrored x n by similar changes in the individual cells. Cohn & Murray also determined the latent periods in their chick heart cultures, and were able to show that, as the embryo grew older, so the cells of its heart took longer and longer to accustom themselves to their in vitro environment. The latent 1Q 7 IR^ ^ ^ , 17 V t Ti 7 / u t J «. '^ t. ^ -e ° 11 2: ^ ^'^ ^7 ^^ in Y ^^ lU / ^^ q ^2 JJ.^ - 4 ,^ ft ~ /^ ^l J' ^^^^ 7 -^/^^"^ ^2 J^^ _• ■^ -( )- "^ ^ "^""^r t^ "^ ^ " _ _ _ Days 4 5 6 7 8 9 10 11 12 13 U 15 16 17 18 Incabation age Fig. 64. Heart-fragments in plasma, ; heart-fragments in plasma + Ringer, H period was defined as the time intervening between the incubation of the cultures and the first appearance of cells protruding from the peripheral margin. As Fig. 64 shows, this value markedly increases, though not until after the 7th day. Unknown to them, Suzuki had previously found just the same relationship between latent period and age in the chick embryo. His figure (Fig. 65) is a beautiful illustration of their conclusions. Growth-rate and latent period are two strangely associated processes. "When older tissue", as Cohn & Murray put it, "i.e. a heart fragment from a 16-day old chick embryo, is placed in a fresh environment, it does not assume immediately a divisional velocity typical of younger tissue. In our experiments the previous growthrate was not approached for 36 hours. In other words, a number of cell-divisions (i.e. generations) were required for it to lose the habit SECT. 2] AND WEIGHT 463 100(— of slow growth previously imposed upon it by its organised environment. This happening is an expression of a process of rejuvenescence or dedifferentiation*. The tissue soon assumes a growth-rate determined by its environment, and as long as the environment can be kept relatively stable, that is to say of uniform composition, presumably there will be no further change in divisional velocity." As in the intact embryo, then, negative acceleration of growth is greatest at the beginning of life. On the other hand, the latent period before growth in vitro begins is greatest at the end of embryonic life. It is likely, therefore, that the factors which determine the initiation of growth and those which determine the extent to which it shall take place in a given time are not identical. Olivo & Slavich and Oda & Kamon, subsequently repeated the experiments of Cohn & Murray and as far as heart fragments were concerned, confirmed them in every particular. Nordmann also confirmed them for liver cells. Oda & Kamon found that the growth-rate in culture was the more rapid the younger the embryo from which the heart had been taken, and the latent period was the shorter. The pulsatile activity of the heart explants was also more hardy the younger the embryo. On the other hand, they did not obtain such clear-cut results with the spleen which to some extent showed a higher in vitro growth-rate the older the embryo from which it had been taken. This doubtful point urgently requires re-investigation, for it has become a commonplace of explantation research that growthrate in vitro follows growth-rate in vivo. Moreover, Allen has found other differences between organs ; thus in saline, embryo heart will not grow at all after the 8th day, intestine after the i ith day, and so on.

This does not imply the acquisition of any pluripotence or totipotence. 5 10 15 Age of chrck giving the heart-cells Fig. 65. 464 ON INCREASE IN SIZE [pt. iii The experiments of the Cohn & Murray type were precisely the converse?of those of Carrel & Ebeling who found that the growthrate of a standard 2-year-old strain of fibroblasts was affected by the age of the hen from which the plasma culture medium was derived. The older the donor the lower was the growth-index, the younger the donor the higher, i.e. the faster the explant grew. Thus the growthpromoting factor must be present both in the cell-protoplasm and in the circulating blood plasma or "milieu interieur". This growthpromoting factor was for long believed to be a hormone or even a vitamine, but it now seems to be much more probably a special collocation of the right nutrient substances, probably protein breakdown-products. Following upon the pioneer work of Ross Harrison, it was found by Carrel; Carrel & Burrows; Fischer; Ebeling; and Carrel & Ebeling that tissues would grow indefinitely in vitro on a medium composed of one part adult plasma and one part embryonic tissue juice. It was then soon found that it was the latter constituent which furnished the necessary factors for cell nutrition and multiplication (Carrel & Ebeling and Carrel). From this point there began an extensive series of researches directed towards the identification of the substance responsible for the effects produced by embryonic tissue juice. Whatever it was, it was not species-specific, for duck fibroblasts were found to grow well in chick embryo extract (Fischer), and vice versa (Kaufmann), while rat tissues could be grown in chick embryo extracts (Mottram), and rabbit embryo extracts would stimulate chick tissues (Carrel & Ebeling) or human tissues (Timofeivski & Benevolenskaia) . Results which are interesting in this connection are those of Loisel who substituted duck egg-white for the natural egg-white of the hen's egg and found that the normal development of the intact chick embryo up to the 4th day was not interfered with by these conditions.* Carrel & Baker examined the fractions obtained from chick embryo juice, and concluded that the growth-promoting factor was associated definitely with the protein part. The precipitated proteins of the embryo juice (acetone, carbon dioxide, alcohols, acetic acid, etc.) never showed a greater growth-promoting power than the original juice, but often nearly as much. Chemical examination showed that the protein in question was a mixture of nucleoprotein

Bouges also, by an injection technique, has replaced the yolk by yolk from other breeds or even species, without interfering with normal development. SECT. 2] AND WEIGHT 465 and a glucoprotein rather like mucin, but when these were prepared in a pure state, they showed no growth-promoting action in tissue cultures. Carrel & Baker tried a great variety of other substances, especially proteins, such as pure egg-albumen, but always without result. Ether extraction, they found, did not remove the growthpromoting factor from the embryo juice proteins, so that it seemed unlikely that it could be of lipoid al nature. Some earlier experiments of Carrel & Ebeling had shown that although pure amino-acids produced a slight stimulation of growth, causing greater cell-migration, they did not produce an increase in the mass of the tissue, as the protein fraction of embryo juice certainly did. This was confirmed by Carrel & Baker, who dialysed the embryo juice in collodion bags of high permeability, or else ultrafiltered it. They found that the dialysed juice did lose a certain amount of its growth-promoting power, but they put this down to denaturation of the proteins, in view of the fact that the amino-acids separated in this way were practically without action on the growth of fibroblasts in culture. Nor was there an enzyme in the juice which produced amino-acids in the dialysing mixture more rapidly than they could diffuse away. It is true that the liquid surrounding the collodion bag and containing the diffusible amino-acids had a slight effect on the increase in area of the explants, the rate being 1-28 times faster than in Tyrode solution, as against 10 times or upward for the protein fraction. This small effect was put down by Carrel & Baker to migration stimulus, and not to increase of mass of the explant. When Carrel & Baker tried to restore the loss of activity of the protein part by adding the amino-acids which had dialysed away, they found it was impossible to do so — a fact which supported the view that the proteins had been denatured by the dialysis. Thinking that the concentration of amino-acids in the untreated juice might be too small to show the presence of the growth-promoting factor, they hydrolysed the juice protein with acids and with trypsin, but the large concentrations of amino-acids so produced proved to be toxic for the cells of the cultures. This toxicity had already been noticed by Burrows & Neymann. Wright also worked with diffusates of embryonic tissue juice and observed a perfectly equal number of mitoses in explants of chick embryo heart, whether the medium used contained embryonic tissue juice itself or the diffusate from it, e.g. 35 ± 5 in the former case and 32 ± 5 in the latter. In another N E I , 30 466 ON INCREASE IN SIZE [pt. iii instance the number of mitotic figures was: saline 2, extract 24, diffusate 26. These results were interpreted by Carrel & Baker as being due to a stimulatory action of amino-acids on migration, and they stated that they observed similar phenomena themselves with pure amino-acids, which, however, gave nothing approaching that large extension of area seen in extracts with the protein fraction. All amino-acids, according to Carrel & Baker, stimulate cellmultiplication and migration, without increasing the mass of the tissue, i.e. without causing growth. In a succeeding paper Carrel & Baker put to the test the suggestion that it was not the proteins as such in the embryonic tissue juice which carried the growth-promoting factor, but rather their larger split-products such as proteoses, broken off from them by the cells of the tissue culture themselves. They found that if tissue juice from embryos was digested with pepsin for 16 or 32 hours, the hydrolysate was rather toxic for the fibroblast explants, but if the digestion was only carried on for 3I hours, the presence of the growth-promoting factor readily revealed itself, growth proceeding seven times as fast as in Ringer solution. The large protein split-products which were evidently responsible could be obtained just as well from peptic hydrolyses of egg-albumen, and commercial fibrin, the latter especially proving a rich source of growth-promoting activity. Chemical examination of the different fractions of the digests, compared with tissue culture examination, showed that the substances responsible were undoubtedly proteoses. Pepsin digests of rabbit brain were found to be active. Commercial "peptones" were found to vary in activity, according to the proportion of higher split-products contained in them. Tryptic hydrolysis was not so satisfactory. Carrel & Baker suggested that the function of the proteose is to furnish a higher concentration of amino-acids to the cells than could be obtained even from their saturated solutions, and to supply them to the cells in this tightly packed, yet soluble and diffusible form. Carrel & Baker's identification of proteoses as the important substances was confirmed by Fischer & Demuth and by Willmer. All these facts led Carrel & Baker to conclude that there was no justification for speaking of a "growth-promoting hormone". The growth-promoting factor is probably no more than a right conjunction of nutrient materials and the appropriate capacities for SECT. 2] AND WEIGHT 467 making use of them. The particular position taken by unhydrolysed embryo tissue juice is of course very important from the embryological point of view, and merits much further research. Still later, Carrel & Baker went on to investigate the growthpromoting activity of digests of pure proteins, fibrin, egg-albumen and edestin. The proteins themselves, of course, had, from the earliest days of tissue culture, been known to be more or less inert (Smyth; Swezy). They now found that pure fibrin, when taken to pieces by pepsin into its proteoses, had as powerful an effect as commercial fibrin, so that the effect was not due to impurities such as blood corpuscles. The split-products of pure egg-albumen, however, did not show so large a growth-promoting action, and Carrel & Baker found that it could be increased by the addition of pure glycine and of pure nucleic acid. It was interesting that digests of vegetable proteins such as edestin and gluten, showed a marked growthpromoting activity. Carrel & Baker also showed that a medium containing a maximum growth-promoting activity could be prepared by digesting calf liver or pituitary with pepsin to the right extent, and that a- and ^-proteoses were equally effective. If now we go back some years we find that in 1921 Carrel & Ebeling found that adult serum strongly inhibited the multiplication of fibroblasts and epithelial cells in tissue culture. As already mentioned, it was not long before they found that this property clearly increased with the age of the animal providing the serum (Carrel & Ebeling). Then in 1927 Carrel & Baker showed that this action was due to changes with age in both the lipoid and protein fractions of the adult serum, and that in all probability it was related to the increase of antitrypsin : a significant finding if the hydrolysis of the proteose by the explanted cells is the most fundamental mechanism involved. Carrel & Baker showed that the growth of fibroblasts was 1-56 times as good in the young as in the old serum. Brody, who was interested in this property of serum as an index of senescence, constructed a chart showing the decline of its growthpromoting power with age, and this is reproduced as Fig. 66. Duration of life of fibroblasts grown in adult serum, plotted against age of fowl giving the serum, declines with a k (instantaneous per cent, growth-rate) of o-i8 and halves itself in 3-9 years. The data from which this graph is derived were given to Brody by Carrel. 30-2 468 ON INCREASE IN SIZE [PT. Ill After these demonstrations of the nature of the growth-promoting factor, the assertions of other workers that a hormone is in reaHty concerned do not seem very convincing. The paper of Heaton may be consuked for experiments done from this point of view. The fact that tissue cuhures of very early chick embryos are not successful and the fact that observers are agreed on the necessity of keeping the cells in numbers for successful cultivation led Wright to investigate how it is that in nature the early stages of the chick are ever successfully passed through. Yolk from the hen's egg proves itself inactive when added to tissue cultures, probably because, as Carrel & Baker showed, yolk-lipoids inhibit growth in explants, but Wright found that on dialysing the yolk a substance passed through the membrane ^ which behaved very similarly to the i; dialysates from embryo tissue juice. 5 Thus in one experiment where the -^ yolk diffusate from 7-8-day eggs was — tested on heart fragments of from E 10- 1 1 days' incubation, the number {£ of mitoses in the treated explant was "S from 1 14 to 127, while in the control «^ it was only 14. Willmer has discussed 160 140 120 100 80 60 these observations in an interesting ° o 40 20 V \ \Duration of life of \ kfibroblasts in serut n N »\ i. 1 s ,0 X, ^ ■^L? ^ r«  ^ Age Fig. 66 6 n Years 10 review. Fischer has elaborated a theory of J^ "desmones" or mutually stimulat- q ing substances, and Burrows has introduced the terms "archusia" and "ergusia" for similar ideas. It is not possible to review the work of these authors here. Burrows & Jorstad did, however, point out that the growth of embryonic cells in tissue culture depends on sufficient crowding and a certain stagnation of the medium. For reasons not at all obvious to the biochemist. Burrows & Jorstad identified archusia with vitamine B and ergusia with vitamine A. Still less satisfactory is the work of Carnot, of Roulet and of Carnot & Terriss who affirm, on very slender evidence, that wounds of metazoal animals heal much quicker when treated with embryo extract than when untreated. Carnot & Carnot have also maintained that injection of foetal extracts into SECT. 2] AND WEIGHT 469 unilaterally nephrectomised animals causes a much increased compensatory hypertrophy of the remaining kidney. Carnot has also studied the effect of chick and rabbit embryo extracts on the growth of tadpoles, using glycerol extracts or the powdered dry substance. There was no influence on metamorphosis, but a great increase of growth, the experimental tadpoles being two or three times as big as the controls. This recalls the older observations of Springer who maintained that the eggs of Arbacia punctulata and Asterias forbesii contained a substance inhibitory to the growth of other eggs of the same species. The retardation was noted in the early rather than the late embryos, and there was a marked tendency for the eggs to stop development when they arrived at the blastula stage. Subsequent work by Peebles indicates that a growth-promoting factor (perhaps a proteose) may also be found in echinoderm embryos. These observations are not very convincing and perhaps the subject might be re-examined with advantage. Experiments which may, when extended further, throw some light on these questions have been made by Skubiszevski, who transplanted chick embryo tissues into adult hens. The complete failures were more or less numerous at all stages of development, and so probably represented errors of technique, but the conspicuous successes gave a curve which was markedly peaked at the 3rd day of development. The grafts grew well without antagonism for as much as 72 days as mesodermal connective tissue, but in three cases (out of 1 78) they resembled a sarcoma at the end of that time. This 3rd day peak may be an important clue. Precisely similar work on grafts was done by Uhlenluth, who transplanted the eyes and skin of larval salamanders into individuals of various ages. These transplanted organs attained maturity not at the time when their original possessor took on adult characteristics, but at the time when their new possessor metamorphosed. They fell into step, as it were, with their new environment. The hormones and growth-promoting substances which regulate such changes are still imperfectly understood, but Babak; Gudernatsch; Swingle; Huxley & Hogben and many other workers, have fully unveiled the importance which thyroxin has in amphibian metamorphosis. Another way of studying the essential nature of the growth-impulse might be to plot the viability for in vitro tissue culture of individual 470 ON INCREASE IN SIZE [pt. iii cells from embryos killed at a definite time. Perhaps it would be found that the younger the embryo, the longer would its component cells remain viable after its demise as a complete whole. Thus Bianchini & Evangelisti found that in guinea-pig and rabbit foetuses the cells of tissues were cultivatable as long as 78 hours after the death of the whole body. And Bucciante, who incubated hen's eggs for 6-12 days and then placed them at o or 15°, found that in vitro cultivation was still possible for as long as 25 days afterwards in the case of epithelial cells, though leucocytes and liver-cells only retained viability for 3 or 4 days. Cells of other tissues occupied various intermediate positions. 2-13. Incubation Time and Gestation Time So far, the total time taken in embryonic growth has not been considered at all. It was indeed said, when the work of Donaldson, Dunn & Watson was being considered, that the act of birth or hatching seems sometimes to have little influence on the course of growth, but as Brody and Schmalhausen have shown, it more commonly aflfects profoundly the rate of increase of size. And it does set a term to the embryonic period (I make no distinction between embryonic and foetal period, as is the custom with some authors), forming at least a convenient index to show whether the growth of the embryo is being accelerated or retarded by external influences. It is therefore important to know the normal incubation or gestation time for as many animals as possible. These are collected together in Tables 60, 63 and 64. It can easily be seen that there is some difference between the estimates of gestation and incubation times in the same animal by different workers, but these discrepancies can only be cleared up by more extended observations. When we come to consider, however, the nature of the law which must presumably govern the length of embryonic life we meet with a remarkable degree of obscurity. Roughly speaking, the commonsense rule that the larger an animal is, the longer its embryonic life must be, is borne out by the figures. Thus in Table 63, where the weight of the adult mammal is seen to vary between 0-014 kilo and nearly 4000-0 kilos and the gestation time between 21 and 600 days, the larger mammals have the longest gestation times, although there are several cases where animals of the same weight have different gestation times (e.g. the pig and the deer) and animals of SECT. 2] AND WEIGHT 471 the same gestation time reach, when full-grown, very different weights (e.g. the antelope and the hippopotamus). However, if the weight when adult is plotted against the gestation time on double-log. paper, a straight-line relation is obtained, individual points not lying very far from the mean. The only reason for using double-log. paper here is the convenience of getting all the data on to the same Table 60. Przibram's figures. Animal Chimpanzee ( 7>o^/o^fe^ m^«r) ... 'M.acacus {Macacus siniciis) Mandrill {Cynocephalus papio) Uistiti (Callithrix jaccus) ... L.emuT (Lemur catta) Lion (Felis leo) Puma {Felis concolor) Ermine {Mustela erminea) Wolf (Canis lupus) Bear (Ursus arctus) Seal {Phoca) German marmot {Cricetus frummtarius) . Mouse (Mus musculus) Rat {Epimys rattus) Rabbit {Lepus cuniculus) ... Guinea-pig {Cavia cobaia) Cow {Bos taurus) ... Sheep {Ovis aries) Goat {Capra hircus) Stag {Cervus elaphus) Roedeer {Cervus capreolus) 'EAand {Alces palinatus) Camel {Camelus dromedarius) Moschus {Moschus moschiferus) Pig {Sus scrofa) Hippopotamus {Hippopotamus amphibius) Rhinoceros {Rhinoceros unicornis) ... Horse (Eguus caballus) Donkey {Equus asinus) Elephant {Elephas indicus) Kangaroo-rat {Hypsiprymnus cuniculus) . . Opossum {Didelphys virginiana) ... Gestation Birth time (days) weight (gm.) 260 1,000 160-210 480 177-210 480 84 33 144 240 105 1,000 92 500 ^4 25 61 250 240 1,500 350 10,000 21 ^5^ 21 1-6 21 5 30 70 63 120 285 37,000 150 4,000 150 2,800 240-270 24,000 280-300 3,000 240-270 60,000 360-400 80,000 160 220 120 1,700 210-250 50,000 510-550 50,000 330-350 40,000-70,000 360-380 20,000 615-628 240,000 u 05 8 05 graph. The line does not meet the co-ordinates at the zero point; it cuts the time scale, no matter what units are taken. This means that no matter how small the mammal, an appreciable time has to be taken in development, and in the case of a mammal as small as a gnat, a surprisingly long gestation time would be observed. Thus the mouse, which is 259,000 times as small as an elephant, does not have an incubation period 259,000 times as 472 ON INCREASE IN SIZE AND WEIGHT [pt. iii short, but only 31-6 times as short. A mammal as small as a gnat, therefore, would probably have a gestation period of 8 or 9 days. The explanation of this must lie in the time requirement of differentiation. lOOOOpr 1000 100 10 TO 0-1 0-01 E =\ V ® «  \ ® ~ \ ® V § • Insectivora © Rodents e \ «  1 ® Even toed ungulates e Odd e Carnivora ® Primates V § Cheiroptera — CO — _o Z5 e \ 9 ■D < ir ® \ — j:: OS 1 '^ C C U3 E § cu 1 c "^ ^ < e e 3 \ ® \ e III ® \ E \° Gest ation-Time per Kilo of Adult i 1 Days \ Mil I 1 1 INI 1 1 Ml!!l 1 1 1 Mill 1 1 Mini K 0-05 0-1 1-0 10 Fig. 67. 100 1000 3000 Another way of looking at the data appears in column 10 of Table 63 and in Fig. 67. In order to eliminate the factor of weight, the gestation time in days per kilo of adult animal is calculated, and this /o 1 00000 / o/ o / o / o A 10000 — o / o /o °/o 1000 100 _ CO -E c c o / / °^ - a; o / AS - S / o o / o 10 / 7 / ° 1 / - CD

/ / 1 1 1 1 1 II! Days Gesta 1 1 1 M Ml bion-Time 1 1 1 1 1 1 1 1 1 1 1 1 1 III 10 100 Fig. 68. 1000 10000 474 ON INCREASE IN SIZE [pt. iii is then plotted on double-log. paper against the weight of the adult animal in kilos. Evidently the relation is also linear in this case, but the interesting thing is that the smallest animals take far the longest time to construct unit weight. Thus the mouse performs the feat of producing a kilo of mice in 1 790 days while the elephant produces a kilo of elephant in 0-16 day. This must be due to the fact that contained in i kilo of mouse there is a great deal more organisation and differentiation than in i kilo of elephant, in other words that the degree of heterogeneity is greater. Whether all mammals can make unit quantity of differentiation in the same time is a question one would like to have answered, but which seems to be at the present time unanswerable. So far, we have only considered the relation between adult weight and gestation time. It would obviously be better to use birthweights for this purpose, but unfortunately only a few are known (see Table 60 taken from Przibram). Nevertheless when the birthweight is plotted against the gestation time on double-log. paper, a straight-line relation is found, except for a slight deviation in the case of the heaviest animals ; this is shown in Fig. 68. The scattering of the points is evidently considerable, but we may say that there is some law which ensures that certain limits shall be held to. Thus if an animal proposes to weigh 100 gm. at birth it must resign itself to an incubation period of between 40 and 1 50 days, while if it is to weigh I gm. it may be between 10 and 30 days in utero. Within these wide limits individual species evidently have the power of making drastic shortenings or lengthenings. Thus gestation time alone may not be a very fundamental constant. In the first place there are great differences in degree of development at birth between such animals as the pig on the one hand and the rat on the other, the former being born almost ready to assume complete motor control of its musculature, the latter by no means ready to do so ; the former able to see, the latter blind ; the former covered with hair, the latter hairless. Any relation between weight and gestation time can therefore only be approximate, and the law governing it must be, as it were, elastic. Again the difference between polytocous and monotocous animals will make itself felt, and the large differences between the relative weights of newborn and mother. The following table (Table 61), which has been constructed from Franck's information, shows how large these are: SECT. 2] AND WEIGHT 475 Table 6] . Total mass Weight of foetal We ght of of one Weight of tissue formed mother new-born mother I : X I : X ^ 1 K J Y Y X X Man 19-1 191 Horse 14-6 14-6 Cow 15-5 15-5 Sheep 12-9 12-9 Dog 7-5 23-5 Cat Rabbit 8-9 8-5 37-3 43-1 Pig ... . 8-2 980 ® opening of eyes ^ appearance of eye-fissure " " rodent teeth And there is also the consideration that gestation time in some animals must be arranged to suit the grazing season. This factor would probably account for a good many of the divergences of species from the line shown in Fig. 68. Again, within the individual species, birth can apparently be shifted to some extent backwards and forwards. Bluhm's work shows that the opening of the eye, the appearance of the ears, and other marks of increasing differentiation in the mouse, occur at a fixed time after conception, so that the smaller the birth-weight the longer the time between birth and the appearance of the mark in question. This relation is illustrated by Fig. 69. What governs the incubation times of birds? The problem has been much discussed, but by far the best treatment of it in the literature at present is the book of Bergtold. Of the 19,000 species of birds known, we have information concerning the incubation periods of 625, and although most of the facts are given in Table 62 Bergtold's book must be consulted for the full material. The length of the incubation period varies more or less with the size of the bird. Fere long ago pointed out that the smaller the egg the smaller the incubation time: thus: Birth weight Fig. 69. Duck Hen Weight of Days egg in gm. Ratio 25 739 I : 084 21 6018 I : 0-815 476 ON INCREASE IN SIZE [PT, III but this strict relation does not hold for many birds and is even more elastic than with mammals. Thus the swift and the raven have the same incubation period in spite of their different sizes, while the kiwi and the hen are very similar in size but have quite 100 10 100 Weight of Adult Bird in Ounces Fig. 70. 1000 different incubation periods. The lapwing, again, though smaller than the woodcock, undoubtedly has a longer incubation period. Nevertheless, when a broad view of the whole subject is taken, and the incubation time is plotted against the adult weight on doublelog, paper, a definite trend does appear (see Fig. 70) and the same 100 1-0 lO'O 100 Weight of Egg in Ounces Fig. 71. kind of picture is obtained when the incubation time is plotted against the egg-weight (see Fig. 71). The most interesting thing to notice is the slope of these two lines, which is in both cases much less considerable than in the mammalian graph of Fig. 68. In other words, if the weight of any mammal is multiplied one thousand times, the gestation period will be prolonged by about ten times, but if SECT. 2] AND WEIGHT 477 the weight of a bird is muhipHed one thousand times, the incubation period will only be prolonged about four times. Similarly if the eggweight is increased by one thousand times, the incubation time is only prolonged four times. No doubt this does not take us very far, but it is always a step forward to have all the information on one graph. Table 62. Bergtold's figures. Weight of Incubation adult bird Weight of Order Species time (days) (oz.) egg (oz.) Struthionidae Ostrich 36-60 4000 48-60 Dromaeidae Emu* ... 56-63 — 20 Spheniscidae Emperor penguin 49 1440 16 Adelie penguin 37 42 4-5 Diomedeidae Albatross 60 224-288 — Phaethontidea Yellow-billed tropic bird 28 14 1-4 Apterygidae Kiwi ... 42 60-65 14-20 Pelicanidae Pelican 28 512 White pelican 29 240 — Ardeidae Great blue heron 28 96-128 Common heron 25-28 64 Loon ... 29 5'7 Black-crowned night heron ... 24 — 1-2 Ibididae Wood ibis 21 144-192 — Anatidae Domestic duck 27 128 Shoveller duck 28 17 Mallard duck ... 26-28 2-8 Pekin duck 30 — 2-3 Grey wild goose 28 160 Greater snow goose ... 29 80-104 — Canadian goose 28 128-224 — Domestic goose 28 — 6 Whistling swan 35-40 192-304 — Cathartidae Calif ornian vulture ... 29-31 320 II Falconidae Gyrfalcon 28 84 — Prairie falcon ... 21-28 22-72 — Western sparrow-hawk 21-28 5 — European sparrow-hawk 29-30 5-6 05 Eastern sparrow-hawk 29-30 4 Honey-eater ... 21 32 — American goshawk ... 28 47 — Western red- tailed hawk 28 48-64 2 Buzzard 28 32-40 — Red-shouldered hawk 28 32-48 — Swainson's hawk 25-28 26-56 — American rough-legged hawk 28 30-33 — Golden eagle ... 25-35 160-184 — Bald eagle 28-36 128-192 — Cracidae Globose currasow 28 114 8 Megapodidae Mallee fowl ... 38-41 — 6-5 Phasianidae Domestic turkey 28 — 3-2 Bobwhite 24 5-5-6-5 Scaled quail ... 21 7-8 —

Haswell gives 84 days and 21 oz. for this bird. 478 ON INCREASE IN SIZE [PT. Ill Table 62. Bergtold's figures (cont.). Weight of Incubation adult bird Weight of Order Species time (days) (oz.) egg (oz.) Phasianidae Grey partridge 24 12-13 — Capercailzie ... 26 184 — Dusky grouse ... 18-24 40-56 — Ruffed grouse 24-28 18-40 — Sage grouse 22 128 — Wild turkey ... 28 160-288 — Guinea-fowl ... 25-28 56 1-4 Ring-neck pheasant ... 24 36 1-2 Golden pheasant 21 20-24 10 Silver pheasant 26 — '■5„ Reeves' pheasant 24 ■ — 098 Domestic hen ... 21 64-80 I •9-2-1 RalHdae American coot 14 16-20 — Otididae Great bustard 28 480 — Charadriidae European woodcock ... 20 8-27 — American woodcock ... 20 5-9 — Common snipe 20 3-8 — Spotted sandpiper 15-16 1-53 — Curlew 30 12-14 — Killdeer 26-28 3-1 0-4 Mountain plover 27 05 Laridae Lesser tern 14-16 2-0 — Common tern 21-23 — 06 Columbidae Band-tailed pigeon ... 18-20 12 — Domestic pigeon 14-18 10 05 Passenger pigeon 14-16 12 — Mourning dove 13-14 4-5-6-0 0-4 Cuculidae Roadrunner ... 18, II — Psittacidae Cockatoo parrakeet ... 21 2-9 — White cockatoo 21 21 — Rose-breasted cockatoo 21 19 — Blue and yellow macaw 20-25 37 — Alcedinidae Belted kingfisher 16-24 5-6 0-45 Strigidae Long-eared owl 21 11 08 Barred owl 21-28 20-32 — Screech owl ... 21-25 4-6 0-6 Eagle owl 21-24 112 — Burrowing owl 21-28 6 — Caprimulgidae Western nighthawk ... 16-18 2-7 0-35 Trochilidae Broad-tailed humming-bird ... — o-i 002 Picidae Hairy woodpecker 14 3 — Downy woodpecker ... 12 I '5 — Williamson's sapsucker 14 1-6 — Red-headed woodpecker 14 2-8 — Lewis's woodpecker ... 14 3-8 — Flicker 11-14 4-3 0-25 Tyrannidae Kingbird 12-14 1-6 0-15 Arkansas kingbird 12-14 1-6 014 Say's phoebe ... 12 09 — Alaudidae Horned lark ... 11-14 1-2 — Turdidae Western robin 14 33 023 Eastern robin 14 006 SECT. 2] AND WEIGHT Table 62. BergtoWs figures (cont.). 479 Order Species Mimidae Catbird Bombycilidae Bohemian waxwing ... Troglodytidae Western house-wren ... Laniidae White-rumped shrike... Hirundinidae Barn swallow ... Tree swallow ... Cliff swallow ... Vireonidae Warbling vireo Red-eyed vireo Sittidae Rocky mountain nuthatch Pigmy nuthatch Corvidae Long-tailed chickadee Magpie Long-crested j ay Crow ... Miniotiltidae Yellow warbler Myrtle warbler Ovenbird Redstart Icteridae Western meadow lark Brewer's blackbird ... Red-winged blackbird Rusty blackbird Bronzed grackle Tanagridae Western tanager Fringillidae House finch ... Arkansas goldfinch ... Pine siskin English sparrow Western vesper sparrow Lark sparrow ... Red-backed junco Spurred towhee Black-headed grosbeak Chipping sparrow Lazuli bunting Incubation Weight of adult bird Weight of time (days) (oz.) egg (oz.) 12-13 1-4 0-06 10-16 2-2 — 10 0-5 — 15 20 — 11-13 14 — 0-05 0-06 12-14 — 0-07 12 0-5 — 12-14 0-07 13-14 12 0-65 038 11-14 0-4 — 17 53 — ■ 17 16-18 4-0 0-6 10 035 0-04 12-13 0-45 — 12 — ■ 0-09 12 — 0-05 15 14 te 0-2 0-18 10-14 I •6-3-0 — 13-16 2-2-5 3-8 12 i-i — 14 066 0-08 12-14 0-47 — 13-14 0-43 — 12-14 I 05 0-09 11-13 0-9 009 12 095 0-07 11-12 0-7 — 12-13 1-5 — 10 1-3 OIO 10-12 — 0-05 12 — 0-07 Gurney's theory was that incubation time depended on longevity. The view of Gadow — at first sight more acceptable — was that the developmental period as a whole was uniform and the longer the Qgg period the shorter the nest period. Yet this simply raises another question, and while it is more difficult to determine the total developmental period than the incubation period alone, the difficulty of relating the preparatory period, whatever it is, to the causal factor, still remains. No doubt the differences between nidicolous and nidifugous birds are removed by this means. Glaus' theory was that incubation period depended on egg-size, i.e. egg-weight, but this 48o ON INCREASE IN SIZE AND WEIGHT [pt. iii as we have seen is only true within wide Hmits, for the ostrich and the kiwi have equal incubation lengths, yet the ostrich's egg weighs 3I lb. while that of the kiwi weighs less than i lb. It is natural that if incubation time depends to some extent directly upon egg- weight, it should depend upon bird body-weight ; for as Huxley has shown, the egg-weight varies closely with the body-weight, though the eggs of large birds are not as large as they should be in proportion. Lastly, Pycraft had a theory that incubation time depended on yolk-weight, but as neither he nor anyone else accumulated any data with which to test the hypothesis, and as it is not in any case a very attractive one, it may be dismissed at once. In Bergtold's view the body-temperature of the parent bird is the important factor. It is likely a priori that the larger the bird the lower its body-temperature, and a degree or two may make a big difference. Bergtold gives in his book a long list of bird temperatures and it certainly seems that the smaller the bird the higher the reading, but unfortunately the data are as yet too few for it to appear whether the exceptions noted above as destroying other theories are abolished on this one. Bergtold's theory is complicated by various taxonomic considerations, in which he supposes, following Sutherland, that the higher a bird is taxonomically, the higher its temperature. As the smaller birds (and mammals) are believed to be the most recent palaeontologically, this may well be the case. In favour of Bergtold's view are the experiments of Heinroth who reported that the eggs of the Egyptian goose hatch in 28 days under a common hen and in 30 days under a Muscovy duck. It is known (see Fig. 83 a) that within narrow limits, the speed of embryonic development in ordinary hen's eggs can be controlled by temperature regulation. "The diminishing size of birds," says Bergtold, "accelerated the metabolic rate, elevated the body-temperature, and so shortened the incubation period." According to Bergtold the scanty data of reptilian incubation times support his temperature theory. Returning now to the comparison between mammals and birds which was raised by Figs. 68 and 71 in which the slope of their weight/incubation-time lines was seen to be different, it is interesting to plot the two on the same graph, as is done in Fig. 72. It can now be seen that not only are the slopes different, but the absolute values are also different, so that on the whole it takes less 31 482 ON INCREASE IN SIZE [pt. iii time to make an equivalent birth-weight of bird than of mammal. (The hatching-weights are here obtained by taking 75 per cent, of the egg-weight in grams, the remaining 25 per cent, being of course divided between shell-weight, weight of membranes left behind, weight of water-vapour evaporated during incubation, and weight of material combusted in the same period.) It is also evident that the largest bird is, as regards birth-weight, 250 times as small as the largest mammal. We are thus left with the following three considerations (which apply wholly to birth-weight) : (i) Although there are mammals as small as the smallest birds, there are no birds as large as the largest mammals. In fact the largest bird is only a little larger than the half-way point on the mammalian line. (2) The time required to make a given weight of bird is always less than that required to make a given weight of mammal, as may be roughly expressed by the following table: Birds Mammals Birth- or hatching Incubation Gestation weight (gm.) time (days) time (days) 100,000 — 600 10,000 — 260 1,000 45 150 100 30 55 10 17 32 I II 14 (3) The prolongation of the incubation time caused by raising the hatching-weight a given amount is not so considerable as the prolongation of the gestation time caused by raising the birth- weight by the same amount. The bird is therefore much more rapid than the mammal in its development, and one may well ask whether this is not an adaptation to life within the egg. In Section 9 and in the Epilegomena the conception of the "cleidoic" egg will be developed, but without forestalling those discussions, it may be said here that eggs such as those of reptiles, birds and insects, with their isolation from their terrestrial environment, quite unlike the close dependence of many aquatic eggs upon the sea, are closed systems, characterised, as it seems, by a definite type of metabolism in which protein breakdown is suppressed and uric acid takes the place of urea and ammonia as nitrogenous waste products. If, then, there are serious problems confronting animals which make their embryos develop in closed SECT. 2] AND WEIGHT 483 boxes, especially with regard to the disposal of incombustible waste, is it not possible that their incubation time would naturally tend to be shorter than that of beings such as mammals which can conveniently excrete their embryonic waste products through the maternal kidneys? It is perhaps justifiable, therefore, to see in Fig. 72 the results of the closed-box system, development inside it being adaptively hastened. It would be very interesting to have parallel sets of data for insects and reptiles, and one might predict that they also would take relatively shorter times than the mammals, but so far I have not succeeded in finding any data from which graphs could be constructed. As for the comparatively mean size of the largest bird at hatching compared with that of the largest mammal at birth, it has been probably more than once suggested that eggs above a certain size would begin to suffer from prohibitive mechanical difficulties. An egg large enough to produce a bird as big as an elephant at birth would require, either internal struts, which would be impracticable, or else an extremely thick shell (see Friese's work, p. 239) which would raise great difficulties with respect to gaseous exchange. It is likely, therefore, that 100 days is the extreme limit to which oviparous animals can prolong their incubation time (without hibernating), as against the 600 or more which are possible to mammals. Is this connected with the extinction of the Aepyornis? It might well be asked at this point how it was that the extinct reptiles attained their prodigious size if they were oviparous, and the answer seems to be that for the most part they were not. Some form of ovoviviparity was common, judging from the numerous finds of small skeletons within the abdominal areas of the larger ones. Whether these were really embryos or perhaps rather remnants of undigested food is not yet, and probably never will be, certain, but the question has been discussed by Fraas; Liepmann; van Straalen, and others and the general opinion is that they should be regarded as embryos. We may conclude that the relatively rapid development of birds is an adaptation to cleidoic life, perhaps associated with the high temperatures of birds. It is interesting that hibernation of embryos is not unknown. The best known case of this is probably the silkworm, the embryo of which spends about 8| months in a more or less quiescent state, not advancing to any extent with its development. Dendy reported in 31-2 484 ON INCREASE IN SIZE [PT. Ill 1898 that the embryo oi Sphenodon, the tuatara Hzard, had an incubation period of 13 months, of which something hke 9 were spent in a hibernatory state. Boulenger observed much the same thing in the case of the European pond-tortoise, Emys orbicularis, which has an apparent incubation period of no less than 23 months. Still more extraordinary is the case of some mammals which possess the power, according to Reinhardt and Prell, of hibernating in the embryonic, partly-completed, state (mole, roedeer, bear, badger, pinemarten, and stonemarten) . And as for the insects, Regen has shown that the eggs of a locust, Thamnotrizon apterus, laid in September, hibernate two or three winters and finally hatch out in March. Hibernation, indeed, is very common among insect embryos, the mosquito for instance {Aedes flavescens) occupies 7 months in its ^gg (Hearle). Table 63. Gestation times of mammals. Species s .a c J5 1 3 i-t u E C3 3 i 1 3 c c 1 3 S *-■ CQ S ca c3 "Is Marsupials S3 m ffi 0, iS > Sg3 a Opossum — — — 8 — — 13 • — — Kangaroo-rat — — — 8-5 — — — — Small kangaroo — — — — — 38 — — Large kangaroo — — — 39 — — 40 — — Insectivora Mole — — — — — 29 30 0-283 105-8 Ant-eater — ■ — — 190 — — — — — Hedgehog 49 49 — — — 42 49 0-175 280-0 Cheiroptera Bat — — — — — 34 36 0-028 1280-0 Cetacea Whale — — 365 360 315 — — — — Dolphin — — 300-360 — — — — Rodents Mouse 21 21 — 21-23 — 24 25 0-014 1790-0 Rat 35 35 — 21 • — • 30 35 0-340 103-0 Rabbit 28 28 — 30 — 29 30 1-360 22-0 Hare — 28 — 28-35 — 29 30 3-640 8-2 Squirrel 28 — — — 28 30 0-340 880 Beaver — 119 — 42 — "9 42 13-600 3-1 Marmot — 35 — 42 — 35 German marmot — 28 — 21 ■ — • — — — Guinea-pig — — — 63 — 64 — — Even-toed Ungulates Sheep 147 147 — 150 — 150 150 29-450 5-1 Chamois 154 154 — — — — 140 36-250 39 Gazelle 154 154 — — — — 150 36-250 4-2 Red deer 266 168 — • 240-270 — 280 245 77-000 3-2 Gnu — — ■ — — — — 238 226-200 1-05 Reindeer — — — 280-300 — 232 210 126-500 1-7 SECT. 2] AND WEIGHT Table 63. Gestation times of mammals (cont.), 485 Species £■ >> Even-toed Ungulates 2w Elk — ^ ^ c Llama Camel Antelope Giraffe Ox Bison Roebuck Zebra Goat Ibex Odd-toed Ungulates Horse Ass Pig Hippopotamus Rhinoceros Elephant Zebra Boar Tapir Carnivora Weasel Otter Polecat Puma Leopard Marten Dog Wolf Fox Bear Cat Lynx Panther Jaguar Lion Tiger Badger Ermine Ferret Seal Primates Monkey Baboon Gorilla Man Chimpanzee Macacus Mandrill Uistiti Lemur 168 365 305 154 300 300 "9 505 670 63 70 63 210 56 42 168 305 154 300 "9 35 "63 56 63 70 63 210 56 63 70 — 240-270 — — 330-360 — — 360-400 — — 180-210 — > 266 328 t53 ■ -? 260 300 300 m S «  272-000 1 13-000 181-000 — 315 — 280 — 168 431 249-000 300 590-000 280 816-000 150 150-180 330-350 360-380 120 210-250 510-550 615-628 345-375 120 400 — 151 119 35 63 63 35 63 62 8-160 1-240 92 93 60 63 60 240 56 56 63 92 63 217 50 63 105 105 90 63 63 63 120 56 70 63 — 100 — no — no 63 1-360 22-500 40-700 6-800 135-500 5-430 127-000 113-000 158-000 158-000 74 350 245 280 — — 280 260 160-210 177-210 144 — 2'ia A^ O a 0-96 2-65 1-7 1-7 0-5 03 356 357 114 . 350 365 100 680-000 362-000 81-800 0-5 i-o 1-2 235 658 235 600 2715-000 1810-000 3625-000 0-09 0-3 O-ID 7-7 50-0 66-1 2-8 1-55 9-3 09 10-3 0-5 0-9 0-7 0-7 210 — — 210 22-600 9-3 280 136-000 2 06 280 63-000 4-45 486 Man ON INCREASE IN SIZE Table 64. Gestation time and incubation time.* [PT. Ill Species Albino rat Norway rat " . "* Guinea-pig Sheep ,, ... Cow Horse Opossum Rhinoceros ... Elephant Jackal {Anubis pavian) Dog Pig Goat Coyote {Canis ockropus) ... Raccoon {Procyon californicus) ... Coatis {Nasua ?) Opossum {Didelphys virginiana) ... Chimpanzee Mammals Extreme variation Gestation time in days 280 272-5 272-2 272-2 (calc. from date of last menstruation) 280-5 (conception calc.) 270 279-14 (primipara) 281-99 (multipara) 271-0 282-5 (fo'" male foetus) 284-5 (for female foetus both calc. from last menstruation) 282-8 (for male foetus) 282-0 (for female foetus both calc. from last menstruation) 272-6 (for male foetus) 267-5 (for female foetus both calc. from one definite coitus) 279 21-6-22-64 Gestation time may be prolonged from I to 6 days if it is simultaneous with suckling 21-0 23-5-25-5 (probably suckling) 64 146 145-153 — 275-291 (according to breed) 210-335 334-359 264-420 13 — 540 — 630 — 210 — 60 55-68 120 104-133 151 — 65 65 71 13 210-245 137-162 Birds Investigator Hippocrates Leuckart Lowenhardt Hasler Schlichting Zollner Ahlfeld Robertson — Siegel t Hecker Stotsenberg King Lantz Miller Draper Schwarz Sabatini [Brody Franck-Albrecht and Ewart Heuser & Hartmann Feldman Heinroth Schwarz Asdell Gander — Borland & Hubeny Species Swan Goose Duck Hen

See also Tabulae Biologicae, vol. 6. I Siegel's figures were obtained during the war on the basis of very accurate data (see also Jolly) . Incubation time in days Investigator 42 Davy 35 j> 28 >} 21 „ SECT. 2] AND WEIGHT 487 Table 64. Gestation time and incubation time (cont.). Birds (cont.) Incubation time Species in days Investigator Turkey 28 Davy Guinea-fowl 30-31 ?» Partridge ... 27 Pheasant 23 Red grouse 23 Pigeon 14 Turtle-dove 14 Canary 13 Wren ••• ..• • 10 Martin {Hirundo urbica) 12-13 Meyer Swift {Hirundo apus) 16-17 Eagle-owl [Bubo maximus) 21 Goshawk (Astur palwnbarins) 21 Sparrowhawk {Accipiter fringillarius) 21 Stockdove {Columba oenas) 17 Turtledove {Columba turtur) 16-17 Pheasant {Phasianus colchicus) 24-26 Cock-of-the-wood {Tetrao urogallus) 28 'Rldick gTonse {Tetrao tetrix) 21 Vheasant {Tetrao perdix) ... 21 Swan {Cygnus olor) 35-42 Wild duck {Anas boschas) ... 29 Humming-bird 12 Milne-Edwards Hen 21 Duck 25 Cormorant ... 25 Guinea-fowl 25 Turkey 27 Goose 29 Peacock 31 Swan 42 Cassowary ... 65 Turkey 29 Evans Guinea-fowl 25-26 3 J Duck 28 33 Partridge ... 25 J3 Humming-bird 12 Schenk Calcutta hen 27 jj Peacock 31 3> Stork 42 )5 Cassowary ... 65 J> Goose 29 » Duck 21 J» Guinea-fowl 21 33 Blackbird {Turdus merula) 15 Evans* 'Redstart {Ruticilla phoenicurus) 14 j> Robin {Erithracus rubecula) 13-5 )» Sedge-warbler {Acrocephalus phragmitis) . 15 j> Great tit {Parus major) 14-9 » Pied wagtail {Motacilla lugubris) ... 14-5 >9 Swallow {Hirundo rustica) ... 155 » Skylark {Alauda arvensis) ... 135 » Common tern {Sterna fluviatilis) ... 21-5 » Redshank ( Tetanus calidris) 23 »> Stormy petrel {Proc ellaria pelagica) 36 3

According to Evans the data in Giglioli's report are not reliable. 488 ON INCREASE IN SIZE [PT. Ill Table 64. Gestation time and incubation time (cont.). Species Condor {Sarcorhamphus gryphns) ... Buzzard {Buteo vulgaris) Turkey Hen Duck Pigeon Rhea {Rhea americana) Buzzard {Buteo buteo) Kestrel {Cerchneis tinnunculus) Pheasant {Phasianus colchicus) Nightingale {Turdus philomelos) Yellow-hammer {Emberiza citrinella sylvestris) Chaffinch {Fringilla coelebs) Goldfinch {Acanthis cannabina) Robin {Erithracus rubecula) Hedge-sparrow {Prunella modularis) Lesser whitethroat {Sylvia curruca) Sedge-warbler {Acrocephalus phragmitis) ... Golden vulture {Gypaetus barbatus) Common pigeon: ist egg of the clutch ... „ ,, and egg of the clutch... Birds (cont.) Incubation time 54 days 31 26-29 „ 19-24 „ 29-32 „ 17-20 „ Marine turtle {Thalassochelys corticata) Python {Python molurus) Loggerhead turtle {Caretta caretta) Green turtle Loggerhead turtle {Caretta caretta) European tortoise {Emys europaea) Tuatara {Sphenodon) Crocodile ... Alligator Python {Python reticulatus) Python {Python molurus) Black snake Fox snake ... Corn snake Yellow rat snake ... Ring snake Milk snake ... King snake Coral snake Dogfish {Scyllium catulus) ... ,, {Scyllium canicula) Ray {Raia batis) Herring Sturgeon {Acipenser stellatus) Plaice Flounder Trout {Salmo fario) American flounder {Pseudopleuronectes americanus) Japanese salmonoid {Plecoglossus altivelis) Pike Reptiles 29 30 28-5 „ 25 13-5 » 12-5 „ 13 II 14 12 II-5 » 12-5 „ 53 16-42 „ 16-89 » 47 days 2i months 64 days 8^ weeks 8i „ 20-44 „ 52 12 8i „ 6-8 „ 10 „ 8* „ 74-8i „ 6-8 „ 11 „ 6 „ 8 „ 6-8 „ Fishes 157-178 days (av. 1 69) Bolau .. 234-280 ,, (av. 261) Investigator Broderip Anonymous Vignes Rozanov Groebbels & Mobert Schumann Cole & Kirkpatrick Tomita Cunningham Hildebrand & Hatsel Bergtold Dendy Bergtold Detmers Bergtold 9-10 months 12 44-80 hours 6-15 months 3-7 205 days (2°) 82 „ (5°) 41 „ (10°) 26 „ 10-25 days 9 Beard Hoflfmann Derjavin Dannevig Haempel Scott Nakai Kvasnikov SECT. 2] AND WEIGHT Table 64. Gestation time and incubation time (cont. Crustacea Species Wa.teT-t\ea. (Daphnia pulex) Lobster Amphipod {Gammarus chevreuxi) ... 489 Incubation time 75-85 hours ID months 8-9 days Insects Cockroach (Periplaneta orientalis) ... ... 75 days Tent-caterpillar moth {Malacosoma americana) 9-5 months Sarcophagidae ... ... ... ... Less than a day y{ous,e ^y {Musca domestica) ... ... 8-12 hours Lepidoptera ... ... ... ... Several months Phasmidae ... ... ... ... 2 years Ant {Aphaenogaster fulva) ... ... ... 17-22 days ,, {Myrmica rubra) ... ... ... 23-24 ,, Dragonflies ... ... ... ... 3 weeks Oriental peach-moth {Laspeyresia molesta) 4 days Butterfly [Diacrisia virgimca) ... ... 5-5 ,, ^■win&-\ouse {Haematopinus sui) ... ... 14 ,, Beet Leafhopper {Eutettix tenellus) ... 10-50 days Stonefly {Perlodes mortoni) 91 days ,, (Ptrla carlukiana) ... ... ... 59 ,, ,, {Nephelopteryx nebulosa) ... ... 20 ,, Mayky {Siphlurus armatus) 15 weeks ,, [Ecdyurus venosus) ... ... ... 15 days Molluscs West African land-snail {Achatina variegata) Periwinkle {Littorina littorea) Worms Roundworm {Ankylostoma duodenale) 40-60 days 6 days -3 days Investigator Ramult Allen Ford & Huxley Zabinski Rudolfs Imms Fielde Needham Snapp & Swingle Johannsen Weber Severin Percival &Whitehead Pycraft Haves Looss Table 65. Time-relations of early development. Amphibians Wilson's figures (15° C.) Species Salamander {Amblystoma punctata) Frog {Rana temporaria) ... Small wood-frog {Chorophilus triserratus) Hertwig's figures (14° C.) Fertilisation to first cleavage (hrs.) 10 35 1-5 Cleavage periods (min.) 1st no 75 30 2nd 100 60 40 Rana temporaria Gastrulation Formation of medullary plate Closure of medullary folds Appearance of tail-bud Appearance of tail and gills Appearance of tail-fin Beginning of operculum ... Gastrulation (hrs.) 3rd 100 50 30 Beginning 60 42 End 78 50 13 Days after fertilisation 1-3 4-2 5-5 6-8 8-6 no 13-4 Appearance of external gills (days) 19 14 35 Adult (days) 100 70 30 490 ON INCREASE IN SIZE AND WEIGHT Table 65. Time-relations of early development (cont.). Molluscs A. Richards' figures: [PT. Ill (Opisthobranch) Planorbis Time Intervals from in fertilisation minutes Haminea virescens , ' , Time Intervals from in fertilisation minutes Fertilisation ... o 1st cleavage ... 90 2nd cleavage ... 165 3rd cleavage ... 240 4th cleavage ... 315 5th cleavage ... 375 E. G. Conklin's figures: 90 75 75 75 60 ASCIDIAN o 85 136 195 245 285 Cynthia 85 51 59 50 40 Fertilisation 1st cleavage 2nd cleavage 3rd cleavage 4th cleavage 5 th cleavage 6th cleavage 7 th cleavage 8th cleavage (218 cells) Hatched tadpole ... Time Intervals from in fertilisation minutes o 40 70 90 no 130 150 170 190 310 40 30 20 20 20 20=" 20 20 120

Beginning of gastrulation. The inner significance of the length of embryonic life relative to the life-span is most obscure. Some interesting remarks have been made by Moulton, who has pointed out that over the whole lifespan, the chemical changes are much more intense in the earliest periods, i.e. pre-natal and to a certain extent post-natal. This reminds If -a '^ 1 ' h to c ^ C- CZ3 O CD O Z 2 < o it; 2: oo '^t- CO CM ^'"H*;; hb L^ n o c Q) z o 1 o X. -z. i V ■ ^^v^ <L<1 y -,<-Qcr^ BOX-, qq»^i tJ cp lo cp ura CD ir: CO CNl eg ^ ^ ci CD - 1 ] 1 ^- i r 1 1 ■D < ho jaqsM queo j^b^ 492 ON INCREASE IN SIZE [PT. Ill US that Murray's law only applies to the embryonic period, and must not be extended beyond it (see p. 548). A typical graph is that shown in Fig. 73 where the composition of the whole human being as regards water, nitrogen and ash, is considered. The point of sudden cessation of intense chemical redistribution was termed quite logically by Moulton "chemical maturity", and this point, he found, bore a fairly constant relation to the total life-span, as appears from the following table : Table 66. Moulton's figures: Species Man Cow Pig ... Guinea-pig Dog Cat ... Rabbit Rat ... Mouse This, however, was the only relation which did show any constancy, and very little can be deduced about an animal if only its gestation time, or conversely, only its average length of life, is knov/n. Even its composition at birth is not related simply to any of the other variables. These facts, to which Moulton was the first to draw attention, illustrate the truth of the statement just made, namely, that the act of birth or hatching is a comparatively unimportant one in the life of the individual. Probably the time at which it takes place in the life-span has been much involved with the adaptations due to different modes of life, while yet the underlying process of physicochemical maturation has remained unaffected. Extremely few researches have been done on these problems. In 1926 I estimated the non-protein nitrogen in a variety of bird embryos, of different incubation periods, with a view to ascertaining whether the rhythm of chemical differentiation went on at a constant rate in all cases. The curve shown in Fig. 74 is from the data on non-protein-nitrogen of White Leghorn embryos. Similar data were obtained for two other races of domestic hen (21 days), the xt J2 •< 0. a> C «  > < tion age at maximum of 3rd growth cycle (Brody & Ragsdale) (days) 60 ,_, •3 ^ .2 " g"S ut» e Part of life-span passed at chemical maturity % of total life-span Constitution of body at birth A ^0 1 0? a 'v u o5 < 285 80 5300 1285 4-4 82 14 3 285 25 850 435 4-6 76 18 4 120 20 200 345 4-6 82 13 3 64 7 145 114 4-6 78 17 4 61 17 261 4-3 82 14 3 60 1 1 — 160 39 83 13 3 31 10 185 — 84 13 2 25 4 86 75 4-5 88 10 2 20 4 62 86 II 3 SECT. 2] AND WEIGHT 493 pigeon (i8 days), the guinea-fowl (25 days), the duck (26 days), the partridge (27 days), and the turkey (28 days). The results were related to weights of embryos equal to those of the chick embryo on different days, and as can be seen, the values for the other embryos fell uniformly on the chick curve. The length of incubation would thus appear to be a tune played as it were, " adagio " in the turkey and "allegro" in the pigeon.* o EI D m H6N : MuK.ttU^Ko^ Partridse: Hen : \NKUre Wyo>«Ulte Pigeon DUCK Turkey guinea-fowu Hen • Ba/»*»««ei<Ur % of total incubation time Fig. 74 The possible adaptive significance of incubation time has been shown remarkably in a paper by Friedmann, who found that the incubation times of the cowbirds, such as Agelaioides, Molothrus and Tangavius, varied according to their degree of parasitism. M. afer, which is very parasitic, has an incubation time of 10 days (the

For this it was necessary to assume that the pre-natal growth-curves for wet weight were alike, an assumption which was subsequently shown to be legitimate by Kaufmann. The instantaneous percentage growth-rates of the pigeon and the hen do not begin to differ until after hatching ; and this holds, according to her, for heart, liver and eye, as well as for the whole body. The cells of the pigeon embryo, however, are smaller than those of the chick by about 30%. 494 ON INCREASE IN SIZE [PT. Ill shortest known), M, bonariensis, less parasitic, takes 11-5, and M. rufo-axillaris , still less so, takes 12-5 to 13 days. No consideration of the questions involved in incubation and gestation time would be complete without mention of the work of Rubner. Rubner found that, if a graph is constructed having the durations of pregnancy of different groups of animals as abscissae and the respective birth-weights of their young as ordinates, the resulting curve is quite smooth and regular. These facts, which have already been discussed, are shown in Fig. 75, taken from Rubner. A glance shows that the only exception among the animals which Rubner chose is man, who develops very slowly, and does not attain at birth more than a quarter of the weight he should if he resembled other animals. Rubner considered the anthropoid apes to be more like animals than man in this matter, but an observation of Heinroth's makes this doubtful. It has often been pointed out that small birth-weight is probably an adaptive feature of considerable advantage to the human female. The actual figures are as follows : Table 67. Period in days during which Gestation the new-born period Weight of doubles ts weight in days Weight of new-born ,.

■ "^ (Thiel; maternal Weight of in % of (Abder Landois; organism new-born maternal Animal (Bunge) halden) Khmmer) in kilos in kilos weight Horse 60 60 340 450 500 I I'D Cow ... 47 47 285 450 35-0 8-5 Sheep 12 15 154 50 3-9 7-8 Man... 180 180 280 55 30 55 Pig ... 16 14 120 80 2-4 30 Dog ... 8 9 63 — 2-0 Cat ... 9 9 56 — — . — Rabbit 6 6 28 — — — Guinea pig — — 67 0-62 0-087 14-2 Mouse — — 21 0-02 0001 7 8-5 Average 7-6 Rubner studied the heat production of the new-born animals, and found that the larger the birth-weight the smaller the number of Calories put out by i kilo of body-weight per day. Thus for a newborn animal weighing 50 kilos, i kilo of body- weight would produce 34*2 Calories, for one weighing 25 kilos the production would be 42-6 Calories, for one of 12-5 kilos it was 60 Calories, and for one of 6-25 kilos it was 66-6 Calories. The explanation of this obviously was that, since no supply of energy can be built up in toto into SECT. 2] AND WEIGHT 495 potential energy stored in the tissues, but must to some extent be wasted in upkeep metabolism, a wastage would be expected in the case now under consideration, and as the relative surface of the newborn is larger the smaller it is itself, this wastage would be expected to be greater. Rubner then assumed that the heat production of the embryo was about seven-tenths of the new-born animal, presumably on the basis of premature birth figures, but also on 50 ffet r 25 14 A / ?7 / / ffO / / 2(7 19 18 17 16 15 n 13 12 17 10 9 8 7 / / f f^inc y y k / ^ 30 / •s <| / / t ^ / t f ^ ^ y / / §■ cji / / / ^ 20 4 / / / / Sj } / / a "^ / / / i^ A / / / / / 10 d y / (r Tnf 4 p M*^ 9 M nv 1 4 6 f y y\ r^ ,S7 ^M V y Z 1 2 ^ nrp '-h mC' jdi veil ^ 1 <: 4 $ I 5 ' t i 1 or 11 ? 1 3 J i 1 5 1 6 1 7 1 > 7 9 £ V ■? 7 2 Z t i 2 4 ^ 5 ^ i 2 7 A i 2 9 3, 7 3 1 3 2 J } s

Fig. 75 theoretical grounds drawn from the embryo's existence in a thermostatic hydrosphere unaccompanied by much muscular or digestive exertion. This assumption seems a dangerous simplification, in view of the peak in basal metabolism which we now know to exist in some animals before and in some animals after birth. However, Rubner calculated the intra-uterine heat production as follows : Birth-weight of animal (kilos) 50-0 25-0 12-5 625 Cals. put out by I kilo * per day 23-9 298 42-0 46-6

I.e. seven- tenths of the post-natal figures. 496 ON INCREASE IN SIZE [pt. iii Averaging the whole of development, Rubner divided each of these figures by 2, and, taking the duration of appreciable development at six-tenths of the whole (the first four-tenths being a negligible quantity), he multiplied each by six-tenths of the total gestation time in days, thus: Birth-weight of animal in kilos 50-0 25-0 12-5 6-25 Total tion Average no. of Calories gesta- put out by i kilo of time finished embryo 340 250 205 177 x6/io development 204 2631 150 2235 123 2583 106 2470 Average ... 2480 Thus I kilo of an imaginary animal (it would be very like a horse), which weighed 50 kilos at birth, would eliminate 2631 Calories during its whole gestation period, while i kilo of an imaginary animal which weighed 6-25 kilos at birth would eliminate 2470 Calories, its relatively more intense heat output being compensated for by its shorter gestation period. Rubner called this relation the "fundamental law of intra-uterine developmental energy". In the above discussion, we were dealing with imaginary examples, but Rubner calculated out the figures for actual animals, as follows : Average no. of Calories put out by I kilo of embryo throughout Animal Birth-weight development Horse 50-0 2028 Cow 35-0 1915 Sheep 3-9 2728 Pig 2-4 2210 Dog 2318 Average ... 2240 Now if we assess the calorific value of the formed kilogram of finished embryo at 1504 Calories, 2480 plus 1504 Calories, i.e. 3984 Calories, are required for the formation of i kilo of embryo during intra-uterine life in the higher mammals. Rubner spoke of a "growth-quotient" in this connection, defined as follows: Energy stored in the tissues x 100 Energy stored plus energy given off as heat * In the case of pre-natal life, it was 1500/4000, i.e. about 38 per cent. There is a certain contradiction here between this efficiency datum SECT. 2] AND WEIGHT 497 and the conclusions of other workers (for which see the section on energy relations), and it is quite certain that the efficiency changes during development. Rubner's law of constant energy requirement during gestation time does not equate well with the findings of Brody, Moulton and others, but it must be remembered that the "law" rests on a very inadequate basis, nothing more, in effect, than a few rough calculations in which some doubtful assumptions are involved. Rubner emphasised the fact that the value of about 4000 Calories, which seem to be required to make one kilo of tissue during embryonic development, is lower than the corresponding value of about 4800 for post-natal life. From this it would appear that the efficiency is greater before birth than afterwards, but this statement needs much qualification. Rubner also showed that the energy consumed per kilo in doubling the birth-weight is approximately the same for different animals. Thus: Energy consumption per kilo in doubling Animal the birth-weight, Cals. Horse Cow Sheep Pig Cat Rabbit 4>5i2 4^243 3>926 3,754 4>304 4,554 Average ... 4j2I5 Man ... ... ... 28,864 Robertson has shown that the generalised form of this rule would be — = a. log X ■\- h, X where E is the energy-consumption, x the weight of the animal and a and b constants which are the same for all species except man. But we have digressed already too far from the problem of growth. Before taking up the effects of external agents on the growth-rate, it will be worth glancing for a moment at the relations between gestation time and the life-span. Buffon in the eighteenth century and Flourens in the nineteenth maintained that the life-span was five or six times the youth-period (see Lusk), but Weissmann showed that there were too many exceptions to this rule to make it of any N EI 32 498 ON INCREASE IN SIZE [pt« iii use. Rubner's book contains a discussion of this matter, from which it is to be concluded that, at present, there is no rule which covers all the cases. [See also Hollis and Szabo.] Rubner's "second law", in which he laid down that the total amount of heat eliminated from birth to death in all the higher mammals is the same, i.e. about 191,600 Calories per kilo, with the exception of man, for which the value was 725,700 Calories per kilo, has often been criticised adversely, and does not here concern us. But it is to be observed that it contradicts the lack of relation found by Moulton and others between simple gestation time and life-span, in that, according to Rubner, a constant percentage of the total heat eliminated during life has been eliminated at birth, i.e. 2500 out of 191,600 or 1-305 per cent. This proportion does not, of course, hold in the case of man, but will be greater because of his prolonged stay in utero. Our immediate aim must now be to examine the effects of various physical influences upon the rate of growth of the living embryo, for by the aid of such a study one may hope to penetrate further into the fundamental nature of the process. That chemical influences can also exert a great effect on the growth of the embryo is obvious, but they will be dealt with in succeeding chapters, such as the sections on vitamines, and general metabolism. At present the discussion will be strictly confined to the effects of radiant energy (heat and light) on the rate of growth, and all teratological considerations will be kept in the background. 2' 14. The Effect of Heat on Embryonic Growth The accelerating influence of rise of temperature on embryonic growth was known to William Harvey, though in an unformulated kind of way, for referring to differences in gestation time, or what he calls "the diversity of going with Child", he says: "For the same thing befalls them as happeneth to Plants, whose fruits and seeds, do more slowly and seldom arrive to maturity in cold Countries, than do other Plants of the same kind which are in a fat and warme soile. So Orenges in England adhere to the trees almost two whole years together, before they come to maturity: and Figgs also scarce ever arrive at any perfection here, which are ripe in Italy twice or thrice a year. And the like befalleth the fruits of the Womb". But perhaps the earliest quantitative observation of the effect of heat on SECT. 2] AND WEIGHT 499 development was that of Gaspard, who in 1822 constructed the following table : Developmental Temperature time of snail's (°C.) eggs in days 6-8 45 12 38 20 21 Gaspard also worked on frog's eggs. Davy and Coste, both in 1856, published some figures showing the acceleration of development of the eggs of the salmon in warmed water. Most of the early work was, of course, fragmentary, and for various reasons unsatisfactory. Thus Philipeaux in 1871 observed that the hatching time of axolotl eggs was shortened from 25 to 8 days as warmth increased but he did not take the temperatures, and Vernon stated in 1895 that the optimum temperature for the development of the embryos of echinoidea was from 7° to 22°, outside which limits they rapidly became abnormal. Semper found that nauplii of Branchus and Apus hatch out at a temperature of 30° in less than 24 hours, whereas at 16-20° they require some weeks. Again, lobster larvae reared at 23-27° passed the fourth moult in about 10 days, i.e. 3 days earlier than larvae reared at 19° C., according to Herrick. Cuenot gave the following figures for the hatching time of the locust : °c. ... ... 25 20 15 10 Days ... 50 55 60 65 Table 68, partly from Davenport, shows the time taken in embryonic development at various temperatures for certain fishes. As we shall see further in the section on resistance and susceptibility, there are to be found among embryos very obvious adaptations to the environment in which they are to grow ; thus Rauber has shown that the eggs of minnows and salmon which develop during the winter will not grow at all at temperatures much above 12-15°, but will do so at 0°. The critical points below and above which normal development will not go on have been determined for amphibians and birds by various observers, and are tabulated in Table 68. The optimum temperature may be regarded as that at which the smallest number of abnormal embryos are produced ; on each side of it the amount of teratological effect will more or less rapidly increase, while at the same time the rapidity of development will on the one hand be increased and on the other hand be retarded. When a certain 32-2 500 ON INCREASE IN SIZE [PT. Ill Tab •le 68. Days taken in development. Temperature water (°C.) of Cod (Earll) Herring (Meyer and Kupffer) Shad (Rice) — 2- O-O 30-0 — — o- 1-9 2- 39 4- 5-9 6- 7-9 8- 9-9 IO-II-9 325 22-0 160 130 40 II — 13-5 20-23-0 14-20-0 E 7 II 3-5 Early quantitative observations. Time taken to reach definite stage, 38° taken as unity [Rana temporaria) Temperature °C. 20 21 22 23 24 25 26 27 28 29 30 31 Edwards' figures o-oi o-oi 0-02 0-02 0-03 003 0-04 0-07 0-I2 0-15 0-35 0-55 Temperature °C. 34 35 36 37 38 39 40 41 42 Fere's figures 0-65 o-8o 0-72 10 I -00 1-25 1-51 0-0 Lillie & Knowl ton's figures: Increase in length in mm. from 24 to 48 hours after hatching. °C. Rana virescens Bufo lentiginosus 9-1 0-9 4-5 3-0 11-12-9 5-3 5-3 13-14-9 4-3 15-5 15-16-9 — 16-3 17-18-9 9-5 ■ — 19-20-9 19-8 21-2 21-22-9 — — ' 23-24-9 ■ — 41-3 25-26-9 31-5 39-0 27-28-9 40-0 — 29-30-9 47-5 56-8 31-32-9 40-2 55-3 33-34-9 43-5 — Higginbottom's figures. Frog [Ranafusca): Temperature °F. f ■^ Date 60 56 53 51 March 11 Egg Egg Egg Egg 20 Hatch 23 External gills — — — 25 — Hatch — — 27 Internal gills — — — 28 — External gills — — 31 — ■ — Hatch Hatch April 4 — — External gills — 6 — Internal gills ■ — External gills 10 Big tadpole — ^ — May 22 Metamorphosis Big tadpole Internal gills Internal gills Aug. 18 — Metamorphosis — — 28 — — Metamorphosis — Oct. 31 — — — Metamorphosis SECT, 2] AND WEIGHT 501 Table 68 [cont.). Marcus' figures. Sea-urchin {Strongylocentrotus lividus). Days from fertilisation o 7 22 29-5 47-5 555 70-5 80 95-5 118-5 Temperature °C. 9 17-19 22 Egg Egg Egg Blastula Blastula Gastrula Bias tula Mesenchyme Gastrula Blastula Gastrula Pluteus Mesenchyme Pluteus — Beginning of gastrulation Old pluteus — Gastrulation — Gastrulation End of gastrulation Prismatic gastrulation — — Environmental temperatures. Temperattire A °C. Species Minnow Salmon Sea-urchin [Echinus esculentus) Frog (Rana palustr is) ,, (Rana vires cens) Toad (Bufo lentiginosus) Hen {Callus domesticus) Salamander {Amblystoma tigrinus) Frog {Rana fusca) ... Hen [Callus domesticus) Green bug [Toxoptera gram nium) Texan cattle-fever tick [Mar garopus annulatus) Brown-tailed moth [Euproctis anysonhea) Bombycid moth [Samia cecro pia) Mealworm [Tenebrio molitor) Colorado potato-beetle [Lep tinotarsa decemlineata) Pine bug [Dendrolimus pini) Hen [Callus domesticus) Rhea [Rhea americana) Locust [Melanoplus differen tialis) Locust [Melanoplus femur-ru brum)

Atkinson considers that heat is conveyed to the eggs thin sheet of rubber. Minimum Optimum o — 2-5 3 6 25 o 17-22 30 32 38 Maximum 12-15 12-15 42 o — 28 35-39 29 — 20-5 — - 37-6 (ist week) 38-3 (2nd „ ) 39-1 (3rd „ ) 1-65 — 7-5 — II — 11 — 9 — 12 — 14 — - 38-3* - 38-5 18 36 18 36 43 32 32-5 28 Investigator Rauber >j Vernon Morgan Lillie & Knowlton Rauber Lillie & Knowlton Schultze Kaestner Prevost & Dumas Edwards Eccleshymer Hunter & Glenn Hunter & Hooker Sanderson Regener Phillips & Brooks Rozanov Bodine better results in artificial incubation are obtained when the by conduction, not radiation, e.g. by covering them with a 502 ON INCREASE IN SIZE Table 68 [cont.). Environmental temperatures (cont.). [PT. Ill Temperature ° C. A f Mini "^ Species mum Optimum Maximum Investigator Japanese cuttlefish {Omma 12 15 18 Sasaki strephes sloani) Frog {Rana fused) ... — 20 rising to 2^ — Hertwig * Toad (Bufo lentiginosus) — 28 „ 31 — j> Frog {Ranafusca) ... — — 23 rising to 30 King* „ {Rana esculenta) — — 33 34 J) Toad {Bufo lentiginosus) — — 33 38 35 Roundworms : {Toxascaris limbata) 6-7 — 38-40 Zavadovski & Sic {Ascaris megaloeephald) . . . 6-7 — 38-40 9J J> {Ascaris suilla) ... 7-8 — 36-38 {Ankylostoma duodenale) lO 20-30 32 Looss 5> 5J — 21-25 Svensson Trematode of Japanese Bil — 27-5 — Miyakawa harziosis {Schistosomum ja ponicum) Japanese teleost {Hypomesus 6-0 — 17-5 Nakai olidus pallas) Goosefish {Lophius amerieanus) 4-0 lO-O i8-o BerrUl Plaice {Pseudopleuronectes 4-5 — Scott amerieanus) Salmon {Salmo irideus) 3-0 6-0 I2-0 Kawajiri

Hertwig's and King's figures are interesting because Rana fusea and temporaria lay early in spring when the water is often freezing: Rana palustris and esculenta later, and Bufo lentiginosus later still. Giglio-Tos has emphasised the importance of this in ecology. degree of external heat is reached, the mechanism of development will be irreversibly interfered with and the growth-rate will suddenly drop. On the cold side the growth-rate will fall off steadily. It is necessary to distinguish these two types of effect, for it is only the latter that concerns us here. According to the kinetic theory, rise in temperature leads inevitably to an increased vigour of molecular movement, and, as is well known, the extent to which this happens gives a measure of the extent to which the process under examination is physical or chemical. For in the former case the amount of increase in molecular motion will only exert a direct effect in speeding up the process, but, if the possibility of a chemical combination is present to complicate matters, the effect will be much greater. The older workers believed on these grounds that it would be very simple to determine whether the nature of any given "master reaction" in living matter was physical or chemical, but maturer consideration and extended experiments have shown that such determination is SECT. 2] AND WEIGHT 503 attended with very great difficulty. Bayliss has drawn attention forcibly to this. The position as regards the embryonic growthprocess is therefore doubtful, and, although its temperature coefficient has been many times estimated, we cannot yet be certain what the real significance of this is. Nevertheless, the recent researches of Crozier and his school have brought us nearer to a sound judgment upon the matter. 2-15. Temperature Coefficients The older quantitative observations were few in number and not very accurate; we owe them to Fere; Lillie & Knowlton; Higginbottom; Driesch; Chambers; Edwards; Semper; and Bury. Some of them are shown in Table 68. They were too few in number to lead to any well-based conclusions, though they certainly demonstrated the fact that, within certain limits, the higher the temperature, the higher the growth-rate. But the classical paper on this subject is that of O. Hertwig, who in 1898 subjected developing frog embryos to various temperatures. He had been preceded by Baudrimont & de St Ange, who, as early as 1846, had observed the accelerating effect of temperature on the developing egg of the frog. His figure, which has often been reproduced, is shown as Fig. 76, and from it one can easily see that the time taken to reach the seventh stage, for instance, is 16-7 days at 20° but 55-6 days at 10°. The time taken at the lower temperature, therefore, is just 3-33 times as long as that at the higher temperature, so x^° of the van't Hoflf equation Vt = X" will be 3-33, where Vt is the reciprocal of the weight gained at a certain temperature and V {t -\- n) is the reciprocal of the weight gained at n degrees higher temperature. Therefore 10 log a: = log 3-33 = 0-5224; therefore log a: = 0-05224 and a; =1-128. In other words, if m days are taken to complete a certain stage of development at 10°, it will take m x 1-128" days when the temperature is n degrees less for the same stage to be arrived at, D'Arcy Thompson calculated all the values of Hertwig's experiments from this simple exponential formula, and obtained a series of curves convex to the abscissa, which showed fair agreement with those plotted from the experimental observations. 504 ON INCREASE IN SIZE [PT. Ill Here 1-128 is the temperature coefficient for 1° (in the common terminology Q.i) but, as it is usual to employ the expression Q,io ^.s 24.0 23° 22° 21° 20° 19° 18° 17° \&°- 15° 14° 13° 12° 11° 10° 9° 8° 7° 6° 5° 4° 3° 2° 1° Fig. 76. the temperature coefficient, the value for the frog is 1-128^°, i.e. 3-34, or, in other words, the time taken to reach a definite stage at 20° multiplied by 3-34 will give the time taken to reach the same SECT. 2] AND WEIGHT 505 stage at 10°. Thus the velocity of the process is more than tripled by a rise of 10° in the temperature, as would ordinarily be expected in the case of a chemical reaction. Judging from Hertwig's curves alone, one should theoretically be able to increase the speed of the growth to infinity by heating the system up, but this of course is not the case, and Fig. 77 taken from Faure-Fremiet's work on Sabellaria eggs shows how, after a certain temperature is reached, the growth-rate may get slower again, owing to the destructive effects of excessive heat on protoplasm. The subject was thoroughly gone into, apparently at Abegg's suggestion, by Peter in 1905, who made many measurements of the development of echinoderm eggs and calculated out the temperature coefficients of Hertwig's results more correctly. His results were as follows: Table 69. Sphaere ^ chinus Echinus Ranafusca Q.1 2-137 First cell-division 2-040 2-352 2-277 Second cell-division 2-404 2-402 1-758 Third cell-division ... 2-264 2-437 1-956 Gastrula 2-490 2-277 2-509 Medullary plate 3-487 2-177 2-272 Closure of medullary folds ... 3-701 2-625 2-042 Appearance of tail-bud 4-102 1-575 1-830 Appearance of external gill 3-151 1-600 1-609 Appearance of tail-fin 3- 1 38 — 1-831 Beginning of operculum 3-245 — 1-707 — 1-546 2-cell stage ... 4-cell stage ... 8-cell stage ... i6-cell stage ... 32-cell stage ... Blastula First mesoderm First spicule ... Full spicules ... Prisma Young pluteus Old pluteus ... From these figures Peter concluded that the average temperature coefficients of the three embryos were not very different, i.e. Sphaerechinus 2-15, Echinus 2-13, and Rana 2-86. He regarded the similarity between the values for the echinoderms as significant. He observed also that Q,io ^^^ i^ot quite a constant : thus between 2'5 and 14-5° it was 3-28 (average) and between 14 and 24° it was 2-26; or in other words that the development at low temperatures was rather slower than it ought to have been if the same temperature coefficient held for all temperatures. Next Peter noted that Q_^q was not the same for each stage of development, for, on the whole, the average coefficient for the cleavage processes was higher than that for the later events. This rule, however, 5o6 ON INCREASE IN SIZE [PT. Ill only held for the two echinoderms studied and not for the frog, thus: Cleavage stages Later stages Sphaerechinus Echinus Rana ... 2-29 2-30 2-23 2-03 2 -08 3-34 Such a comparison obviously suggested interesting conclusions in view of the fact that the two alecithic eggs behaved in the same way 30' 29 28 27 26 25 24 23 22 21 20 19 ]8 o 16 c14 11 10 9 7 6 5 4 3 2 1 h 20 40 60 80 100 120 2 150 200 Minutes 4 Hours Fig. 77 while the yolk-laden one did not. Peter's final procedure was to set side by side a quotation from van't Hoff, in which chemical reactions were said to be increased 2| times for each rise of temperature by 10°, and the overall average of his own work, i.e. 2-499. Nor did he fail to point out that in inorganic chemical reactions also the QjyQ increased as the temperature was lowered. Subsequent workers found many cases in which embryonic growth followed the exponential rule. Erdmann & R. Hertwig gave figures SECT. 2] AND WEIGHT 507 for Strongylocentrotus lividus, from which Kanitz calculated temperature coefficients agreeing with Peter's. Bialascewicz confirmed the work of Hertwig on the amphibian embryo, but obtained a lower average temperature coefficient, namely, 2-40. For the early development of Strongylocentrotus lividus Loeb found in 1908 an average Qj^q of 2-86, which varied thus with temperature: °c. aio 3-13 391 4-14 3-88 5-15 3-52 7-17 3-27 9-19 2-04 10-20 1-90 12-22 1-74 For the early development of Ascaris megalocephalus, Faure-Fremiet found Ohio's as follows : 0-16 6-25 16-23 393 23-32 I 82 and for that o^ Sabellaria alveolata: °c. Q.10 7-13 2-66 2-l6 16-19 1-28 19-22 1-27 Ephrussi (in 1926) studied the different phases of mitosis in seaurchin and nematode eggs, allotting to each a characteristic temperature coefficient. Thus he obtained the following figures, working always between 18 and 25° C. and calculating for each period separately, not in a cumulative fashion from zero hour. Table 70. n A (from zero hour to disappearance of nuclear membrane, zero being copulation of pronuclei in S. lividus and laying in A. megalocephala B (from disappearance of nuclear membrane to the first appearance of the equatorial plate) A plus B, i.e. the whole of the prophase C (duration of the equatorial plate stage) Z) (duration of the whole of the anaphase) E (reconstitution of nuclei after mitosis) A plus B plus C plus D plus E, i.e. the whole mitotic process B plus C plus D plus E, i.e. from disappearance of nuclear membrane to reconstitution of nuclei after mitosis Strongylo centroius lividus Ascaris megalocephala I 2-31 2-20 ? 1-66 2-33 i-oo 2-07 1-22 1-71 2-00 1-39 1-70 1-33 I -80 I •41 1-45 5o8 ON INCREASE IN SIZE [PT. Ill As Ephrussi says, the correspondence can hardly be a mere coincidence, and in any case the fact that change of temperature has almost no effect on process C must be significant. We do not know whether the effect of increased temperature is greater on the mitotic or inter-mitotic period. Bucciante maintained chick embryo explants at 31° and at 41° and counted no more mitoses during a given amount of growth in the latter than in ^ the former, so he concluded that '~ the effect was the same. Loeb & Wasteneys found 2-3 for the embryonic growth of Arbacia and Bohr 2-9 for that of the snake Coluber natrix. Warburg gave 2-5 for the respiratory increase of Arbacia eggs. Herzog, working on the figures of Dannevig for the hatching time of the plaice egg {Pleuronectes platessa), got a Q^io of 2-5, on those of Earll for the cod egg [Gadus morrhua) 3-4 and 2-3, and on those of Ainsworth and Metzger for two different kinds of trout egg 3-4 and 5-3. Later, Bachrach & Cardot obtained a value of 3-2 for the embryonic growth of the slug Agriolimax agrestis between 6 and 23°, and one of 3-06 for that of the water snail Limnaea stagnalis between 1 1 and 32°. For Ascaris megalocephala Zavadovski got a Q^^q of 2-8, but it was markedly higher at low temperatures than at higher ones. Then later, Zavadovski & Sidorov gave Q^io (18-28°) 3-12 for Toxascaris limbata, 3-12 for Ascaris megalocephala and 3-23 for Ascaris suilla, but 1-45, 1-46 and 0-73 respectively between 28 and 38°. Brown has also published a Q^^q for Ascaris, Berrill for Lophius americanus, Harukawa for the oriental peach-moth, Kawajiri for Salmo irideus, Nakai for Plecoglossus altivelis, Kojiyama for Pagrosoma major. But the older phase of the work on the effect of temperature on embryonic growth could not lead to any definite conclusions about SECT. 2] AND WEIGHT 509 the nature of the Hmiting factor or master process, in view of the fact that Q^^Q itself was not a satisfactory standard. When a physical process of diffusion (Nernst) could have a coefficient practically identical with a chemical process (saponification of ethyl butyrate by baryta, Trautz & Volkmann), both being about 1-3, it was difficult to say what might be producing a given Q,io • But more serious was the fact which soon became apparent, namely, that many cases of embryonic development could equally well be said to be simply linear functions of the temperature. Krogh was the first to find this. In considering the figures which had been obtained by Krogh & Johansen and by Dannevig on the rate of development of fish eggs and the hatching times at diflferent temperatures, he found that the relation with velocity was a straight line and not an exponential curve, and that the relation with time was a curve which was not exponential but was a hyperbola. This is seen in Figs. 78 and 5IO ON INCREASE IN SIZE [pt. iii 79, where the Hnear relationship clearly appears, and the velocity increment is directly proportional to the temperature increment. In 1913-he continued this line of thought by finding that the eftect of temperature on (a) the rate of segmentation in the frog's egg, {b) the rate at which definite later developmental stages in the frog embryo are reached, and {c) the hatching time of a water-beetle, Acilius sulcatus, could in all these cases also be expressed by a straight line. In the case of the frog, as Krogh pointed out, it would have been absurd to express this linear relation in terms of Q^^q , for Q^^q would thus become: 3-5 5-3 5-10 4-1 10-15 30 15-20 2-0 and could hardly be said to be a constant if it was constantly varying. Krogh did not compare his results with those of Hertwig, for the latter used rather different technique, e.g. transferring his embryos and eggs very slowly and by stages to the constant temperature basins. This procedure Krogh did not find necessary, though he observed that the limits of normal development are narrower during the earlier stages than they are later. The linear relationship thus found deviated, however, to some extent below 7°. Exactly the same straight line was found when the velocities at which later developmental stages were attained were plotted against the temperature, though here divergence began as high as 12°. The hatching time of the water-beetle eggs behaved in the same way. Next Krogh took the figures of Loeb and Loeb & Wasteneys for echinoderm eggs already mentioned, and found that they also were best expressed by a straight line when —. ; was plotted ■^ ^ in mmutes against temperature. These results are all summed up in Table 71, which gives the geometrical characters of the straight lines, i.e. their slopes, together with other data about the embryos in question. In a further paper Krogh showed that the relation between the temperature and the carbon dioxide production of the chrysalides of Tenebrio molitor, the mealworm, could also be best expressed by a linear relation, and was not susceptible of description by the van't Hoff formula. At this point the question became a matter of dispute concerning the proper way of calculating the results obtained in such experi SECT. 2] AND WEIGHT 511 ments, Kanitz contested the truth of Krogh's view that the van't HofF equation was inapphcable, on the grounds that the straight Hnes might be the result of two or three flat exponential curves. (Kanitz seems not to have avoided confusion between the time taken for development and the reciprocal of the time taken.) Kanitz plotted the log. of the velocity (log 1000/^) against the temperature, and obtained two straight lines in each case, just as Crozier did afterwards, with kinks between 10 and 15°. He assumed therefore that the Arrhenius formtila could be used and calculated Q^^q from it. Table 7 1 . Effect of temperature on embryonic development. Krogh's figures; Embryo Cod (Gadus morrhua and aeglifinus) Plaice {Pleuronectes platessa) Cod {Gadus merlangus) Plaice {Pleuronectes flesus) „ {Pleuronectes platessa) Sea-urchin {Strongylo centrotus purpuratus) Frog {Rana butyrhina) >J 5! Water- beetle {Acilius sulcatus) Sea-urchin {Arbacia) 2^ a 1 J3 «  «  feicia a!:c3 temp een w Itionsl d temp een w ent w ill J3 M° ^5 S C4 „ a! "H M •M U.S2 ■(J .t;.Q u „ .S-^.2c %% •2^ i-i? esi>° Investi Process £;§-§ >-3 J3:S.g •r! 3 01 iJ «T3 Ml gator Fertilisation to -3-3 12-5 3-5-14 -I-I4 Dannevig hatching Do. -2-1 11-6 6-5-12 jj Do. O i8-8 55-14 — 3) Do. . -1-4 30-3 6-5-12 • 5J Fertilisation to -0-4 13-8 3-5-13 0-14 Johansen & 4-9 mm. larva Krogh Cleavage i i-o 0-99 4-20 3-5-22 Loeb Cleavage 2 — 1-57 — ,, Cleavage i 2-7 0-511 7-21 3-5-22 Krogh Later devel. 6- 1 — 12-21 7-5-25k ?j Fertilisation to 9-4 0-65 15-25 15-28 55 hatching Cleavage i 9-2 I 03 15-28 7-31 Loeb & Wasteneys Kraf ka in 19 1 9 found that in Drosophila the effect of temperature on rate of development could be expressed best by a straight line, but its effect on eye-facet number could be expressed best by an exponential curve. The appearance of both forms in one and the same living material made it unlikely that either relation needed correction. It was more probable that the formulae themselves were inadequate for proper analysis. Other lines of approach were made by entomologists on the one hand and by marine biologists on the other. Thus Reibisch worked over the original results of Dannevig, and concluded that a certain amount of heat was necessary for development and had to be supplied from outside, forgetting that [uj LIBRARY 512 ON INCREASE IN SIZE [PT, III no matter how hot the environment, the egg gets all its energy for development from the inside. Apstein, who noted with Krogh that the time/temperature relation was linear, i.e. the product was a constant, formulated another "day-degree theory", which, however, was no improvement. On the Reibisch-Apstein view, a certain sum of heat, a certain number of temperature units or day-degrees, is necessary to complete embryonic growth. Thus if an embryo hatches after lo days at 12° 120 day-degrees would be said to have been required, so that at 6° 20 days would be required. Krogh & Johansen were easily able to show the superficiality of such a treatment. Williamson's experiments with herring, haddock and plaice eggs also seem to show a linear time/temperature relation, but the variations are large. Elmhirst studied the hatching time of Leander squilla and the development of various decapod Crustacea, and observed a constant time/temperature factor. He also went over the large number of fragmentary papers on the effect of temperature on the development of fishes which exist in the literature, mostly in Fishery Board Reports, and in many cases, though not in all, succeeded in finding evidences of the Krogh straight-line relationship between velocity and temperature. He thought that this emerged more clearly if i -9° were added on to each temperature in order to have the freezing-point of sea water as a basis. But, on the other hand, in many cases, such as that of the cod (Fig. 80, taken from Fulton from various official Reports) the time/temperature curve could certainly be said to have a temperature coefficient (2-3 for 10-20° and 3-5 for 0-10°). Blunk working on Dytiscus embryos, and Blunk & Janisch working on those of Blitophaga opaca, obtained straight lines like Elmhirst for velocity/temperature graphs, i.e. hyperbolas for time/temperature graphs. Sanderson has reviewed the work that has been done on the Fig. 80. SECT. 2] AND WEIGHT 513 effect of temperature on the insect embryo, and has in many cases recalculated the figures of earlier observers, such as those of Abbe : hatching of the Rocky mountain locust [Melanoplus femurrubrum) . Regener : hatching of the Pine bug {Dendrolimus pini) . Hennings : hatching of the Typographic beetle ( Tomicus typographus) . Quaintaince & Brues : hatching of the bollworm {Heliothis obsoleta) . Girault : hatching of the bollworm {Heliothis obsoleta) . Jenne; Hunter & Glenn: hatching of the codling moth [Carpocapsa pomonella) . Jenne ; Hunter & Glenn : hatching of a Braconid parasite [Lysiphlebus tritici) and of the green bug ( Toxoptera gramnium) . Hunter & Hooker : hatching of the Texan cattle-fever tick [Margaropus annulatus) . All these data without exception, together with figures collected by Sanderson himself on the incubation oiEuproctis, Samia, Tenebrio, Leptiwo^arj"a,Afa/(2coj'oma, etc., gave curves closely resembling those of Hertwig for the development of the frog. Thus the Q^^q for Jenne's codling moth curve would work out at approximately 2-8 between 14 and 24°, and that for Samia cecropia would be i-g or so between 20 and 30°. In most cases the usual rise in temperature coefficient with decreasing temperature was shown, and in Samia, for instance, Q^io"^ould be 2-5 between 10 and 20°. It is extraordinary that Sanderson made no calculation of temperature coefficients, but he was particularly interested in the practical application of his work and computed his results according to an empirical day-degree system which, as Martini showed, has no physico-chemical meaning, and even confuses intensity with quantity of heat. It is curious that ichthyologists on the one hand and entomologists on the other should have evolved very similar treatments of time/temperature curves both equally unsatisfactory. The temperature coefficients for these insect curves would work out somewhat as shown in Table 72. Peairs later tried to show that, for certain insects, the curve relating time taken in embryonic growth with temperature was a hyperbola, and he did indeed, over a short range, demonstrate the reciprocal to be a straight line, as, if that were the case, it should be. Peairs paid no more attention to the temperature coefficient question than did Sanderson. Precisely analogous experiments and results were obtained by Blunk working on Dytiscus marginalis and by Bodine on various grasshoppers. NEI 33 514 ON INCREASE IN SIZE [PT. Ill Table 72. Temperature coefficients of insect development. Species °C. Q.IO Investigator Browntail moth {Euproctis chrysorrhea) . . . 16-26 2-0 Sanderson Bombycid moth {Samia cecropia) 10-20 2-5 JJ >> jj 20-30 1-9 ,, Mealworm {Tenebrio molitor) ... 10-20 3-8 JJ ,, ,, 20-30 2-1 JJ Colorado potato-beetle {Leptinotarsa io-20 2-5 JJ decemlineata) JJ JJ JJ 20-30 2-0 ,, JJ JJ JJ 20-30 2-2 Girault JJ JJ JJ 20-30 2-0 Girault & Rosenfeld Pine hug (Dendrolimus pini) 15-25 1-4 Regener Typographic beetle ( Tomicus typographus) 15-25 2-9 Hennings Bollworm (Heliothis obsoleta) ... 10-20 4-4 Quaintaince & Brues JJ JJ 20-30 2-0 ,, ,, Codling moth {Carpocapsa pomonella) ... 15-25 2-8 Jenne Braconid parasite {Lysiphlebus tritici) ... O-IO 2-3 Hunter & Glenn JJ JJ JJ 10-20 33 JJ JJ Green bug {Toxoptera gramnium) O-IO 2-5 )j J) JJ JJ 10-20 33 JJ JJ Texan cattle-fever tick {Margaropus annu 10-20 3-1 Hunter & Hooker latus) JJ JJ JJ 20-30 21 JJ JJ J) 5) 'J 30-40 1-8 JJ JJ Codling moth {Carpocapsa pomonella) ... 15-25 2-7 Hammar Gypsy moth {Porthetria dispar) 5-15 4-9 Peairs JJ j> 15-25 1-9 JJ Locust (Melanoplus dlfferentialis) ,, {Melanoplus femur-rubrum) 23-33 2-5 Bodine 23-33 2-4 ,, Green-striped locust {Chortophaga viridi 23-33 2-8 JJ fasciata) Bachmetiev made some experiments like those of Sanderson and Peairs, and Kanitz reviewed all the older literature in 191 5. Kanitz also suggested that the variations found in the pregnancy time of mammals, and especially man, might be related to the differing body-temperature in individuals, for 2° at a Q^iq of 2-5 would in man make a difference of 10 days. This notion was supported by O. Wellmann, who went through the statistics of pregnancy time in mares and found that, while the average was 326 days from July to September, it was 343 days from March to May. Since the researches of Cobelli; Vos; Congdon; and Sumner have all shown that the body-temperature is raised to a slight degree when the environing atmosphere is warm, the gestation time might very well be expected to show the seasonal rise and fall which it actually does. Vicarelli's thermometric observations support Kanitz's theory. SECT. 2] AND WEIGHT 515 2- 1 6. Temperature Characteristics The older phase of the subject ended, then, in a rather barren doubt as to whether the van't Hoff equation was apphcable to hving processes such as those of the growing embryo, or more correctly whether the time/temperature relation was best expressed by a curve of exponential form or by a hyperbola. In so far as it was applicable, it gave definite information that the limiting factor of embryonic growth was probably chemical rather than physical, but that was all. Snyder's paper of 1908 introduced a new period, that of the use of the Arrhenius equation. This expression is K^ = K^eHk'i) or ^ = eHk-fX where K^ and Kj^ are the velocity constants of the reaction in question at the high and low temperatures, Tg and 7"i, chosen respectively, e the base of Napierian logarithms, 2 the gas constant, and /a the gram molecular energy of activation of the catalyst, i.e. the "critical increment of the active substance" if the reaction is monomolecular, or the sum of the gram molecular energies of the substances if the reaction is bimolecular. The temperature is expressed in degrees absolute. If the velocity constants cannot be calculated, the reciprocals of the time taken to do a definite amount of embryo formation or other work may be used instead, so that the relation becomes : T2- Ti = 10, then -^ = d^^ where ^^ and ^2 ^re the times in question, e.g. 10 days from fertilisation to hatching at one temperature, 20 days at another. When of the van't Hoflf equation. The relation may also be expressed: r log Ko — log K-, a = 4'6l —2 — 2 2 — i. II It was originally suggested by Arrhenius as an empirical description of the facts, but it now has through the work of Rice; Rodebush; 33-2 5i6 ON INCREASE IN SIZE [pt. iii and Thomson, a solid theoretical basis. The quantity fx is called the temperature characteristic in distinction from the temperature coefficient, for it is the calculated "energy of activation" of the active molecule in the reaction governing the slowest process in the complex chain of processes under investigation. It is an index, therefore, of the nature of the catalyst in operation, and should do much more than merely distinguish between chemical and physical processes. Reactions having the same catalyst have the same temperature characteristic. For chemical reactions ju. varies between 4000 and 35,000, and can be represented graphically by the slope of the straight line or lines, if there is a break, relating the log. of the velocity in question to the reciprocal of the absolute temperature. The break indicates that one reaction of the catenary has ceased to be the slowest and another has taken its place. The steeper the slope, the greater the increase of velocity with unit rise, and the higher the value of fju. The log. effect/reciprocal of absolute temperature relation does not always give a straight line, but only if the process that is being measured is irreversible. If an equilibrium effect is functioning, then the relation will be a curve. Finally, fju has nothing to do with the Q^ of the van't Hoff equation, for the former has reference to heat of activation and the latter to heat of reaction. The advantages of the Arrhenius equation over the van't Hoff equation are considerable. The former has a much greater range of values dealing with thousands instead of decimal units. The Arrhenius equation also reveals abrupt breaks in the straight lines relating log. effect to the reciprocal of absolute temperature at which one limiting reaction is supposed to take the place of another ; these are masked by the van't Hoff equation. Then Q,io is not a constant while jLt is, at any rate between certain definite temperatures where the breaks occur. But most important of all, Q^^q values cannot be definitely associated with specific types of chemical reaction, such as oxidation-reduction, hydrolysis, and synthesis, so that all one can hope to find out by the van't Hoff equation is whether the process is physical or chemical, while with the Arrhenius equation one may discover, perhaps, of what nature the controlling change is at any given moment. Arrhenius himself in 191 5 made many applications of this equation to biological processes, and it is interesting that he gives as an average value for segmenting eggs fx 14,100, but without any reference or SECT. 2] AND WEIGHT 517 even any intimation whether this was based on results obtained by himself or by others. The matter is curious, for 14,000 is, as will be seen, not now regarded as a zone occupied by growth processes. But the equation was first applied in physiology by Snyder, who calculated /x for various temperature/time effects in muscle action, and compared them with others from various sources. Thus he found a /x of 12,800 calories for the experiments of Peter with echinoderm egg-cleavage and of 16,600 for those of Hertwig on amphibian eggcleavage. He did not, however, carry the question much further, and it was left for Crozier and his collaborators, in a long series of papers, to work out the temperature characteristic for a great number of processes. Before discussing the findings in the case of embryonic growth, we must outline the groups into which Crozier found he could separate living processes as regards their temperature characteristic. The temperature characteristic of many respiration processes, both in vivo, such as the oxygen consumption of Arbacia eggs, and in vitro, such as the combination of oxygen with haemoglobin, was about 16,600. In many cases an abrupt break in the straight lines would occur; thus below 15° /^ for the oxygen consumption and ciHary activity of mussel gills was 16,000 and above 15° it was 11,000, which latter value held at all temperatures tried for the oxygen consumption of the guinea-pig uterus. On the other hand, the rate of progression of all kinds of invertebrates usually had a /A of between 12,000 and 12,500, and such phenomena as the frequency of the firefly flash, the frequency of invertebrate heart beats, and the frequency of noise production also came in this group. A /x of 12,000 would seem to indicate the predominance of a nervous factor, and of 16,000 the predominance of an oxidation. Other processes gave various values of /u,, e.g. 9240 for rate of movement of algae, from 4700 to 10,300 for protoplasmic streaming in plant cells, and 7900 for rate of pulsation of infusorial contractile vacuoles. These would possibly indicate the appearance of a physical factor, probably diffusion. Rate of nerve conduction gave 8080 and 10,700. It is probable that even a difference between 16,000 and 16,700 is significant, for the former may indicate that the controlling catalyst in the oxidation is iron, while the latter may indicate that a dehydrogenation mechanism is at work; again, the value 11,500, which some oxidations show at some temperatures, is 5i8 ON INCREASE IN SIZE [PT. Ill closely associated with hydroxyl ion catalysis. Glaser suggested that the value of 8000 which he found for a phase of Paramoecium locomotion was associated with hydrolysis. Crozier & Stier found that the [x for various processes in the grasshopper was quite different after decapitation to what it had been before, changing from 7900 to 16,500, and showing that, after central nervous system control had been removed, an oxidation was the slowest reaction. "The velocities of vital processes", said Crozier, "are determined by the velocities of dynamically linked systems of chemical transformation differentially affected by external conditions." Table 73 gives a resume of the commonest [x values and their significances. Table 73. Significance 4,7001 8,000 I 9,000 f 1 0.000 j 11,300 1 2.000 1 I2,500[ 16,000) i6,6ooJ 18,000 20,000 24,000 27,000 Physico-chemical Hydrolysis or diffusion Biological ( Protoplasmic streaming < Rate of nerve conduction (Rate of moving algae Hydroxyl ion catalysis of oxidations Oxygen consumption Breaking of i sulphur linkage Nervous factor Catalysis by iron dehydrogenation Activation of iodine Hydrogen ion catalysis Certain hydrolyses Oxidation, oxygen consumption Growth In 1926 Crozier marshalled all the values that had been obtained for p, in biological processes into a frequency polygon, from which it appeared that the commonest types of fx were roughly 8000, 11,000 and 16,000; above that the peaks or modes were not so striking, for the observations were fewer. He also marshalled all the values that had been obtained for breaks in temperature characteristics at definite temperatures into another frequency polygon, from which it appeared that 15° was by far the most important of these, but that 8°, 20° and 30° were also associated to some extent with abrupt breaks in the values of /x. Less common were breaks at 4*5°, 25° and 27°. That both these polygons are definitely multimodal is a fact which makes the whole position more convincing than it might have been; for, if chance had been operating alone, there would have probably been but one peak. SECT. 2] AND WEIGHT 519 These preliminaries concluded, one can examine the results of Crozier for growth in general and embryonic growth in particular. He found on examining the figures for the growth at different temperatures of Drosophila melanogaster embryos (and larvae) (Loeb & Northrop and Krafka) and of the pupae of Tenebrio molitor (Krogh) that there were two temperature characteristics, 27,000 below 15° and about 10,000 above 15°. This was evidence that whatever the 00034 COOdD 00056 Fig. 81. Temperature characteristics of early development in the frog. A, Krogh's data (first cleavage) ; B, Lillie & Knowlton's data (from first cleavage to disappearance of yolk-plug); C, Lillie & Knowlton's data (from other cleavages to disappearance of yolk-plug) ; D, Krogh's data (later development) . limiting factor in growth was, it was not an oxidation process, a conclusion that was strongly supported by the data of Loeb & Wasteneys and Loeb & Chamberlain for the velocity of segmentation (first cleavage) of echinoderm eggs. This worked out at 41,000 below about 11°, 21,000 between 11° and 16°, and 12,400 above 16°. Krogh's data, again, for the segmentation of frog's eggs, gave 22,600 below 13-5° and 10,200 above 13-5°. In no case, therefore, was a typical oxidation temperature characteristic obtained for the earliest stages of embryonic growth. 520 ON INCREASE IN SIZE [PT. Ill Crozier himself emphasised the view that these facts opposed the theory of metabohc gradients (see Section 3 '8), and, in so far as that theory holds that such tissues as have the highest metabohc rate (respiration intensity) must also be growing the 0.0035 Vj° <abs. 0.00d6 Fig. 82. Temperature characteristics of early development in teleostean fishes. A, By Dannevig's data {Gadus); C, Krogh & Johansen's data {Pleuronectes); D, Higurashi & Tauti's data {Hypomesus) ; E, Higurashi & Nakai's data {Plecoglossus) . fastest, Crozier's facts do oppose it. Other work on the temperature characteristics of growth confirmed that of Crozier; Brown, for instance, found jit's of 28,000, 17,210, 7410 and 19,800 for the adult development of cladoceran arthropods, though in one instance it must be admitted that he got a value of 16,950 {Simocephalus serrulatus) . In a later paper, Crozier considered fully the results of many other SECT. 2] AND WEIGHT 521 investigations on the effect of temperature on embryonic growth. Figs. 81 and 82 show some of the results. Fig. 81 includes the data of Krogh; Lillie & Knowlton; and Hertwig on early development in the frog, and Fig. 82 includes the data of Dannevig; Krogh & Johansen; Higurashi & Tauti, and Higurashi & Nakai on development in various teleost fishes. The regularity with which values are obtained for /x of about 20,000 is remarkable; thus the first cleavage o£Rana eggs has 21,900, the third cleavage to yolk-plug disappearance in Rana has 22,000, the total incubation time oi Hypomesus olidus has 23,700 and of Plecoglossus altivelis 23,000. Blunk's figures for the rate of development of Dytiscus marginalis give 19,300 and of Dytiscus semisulcatus 20,000. Ziegelmayer's figures for the rate of development of Cyclops give 15,700, which is on the low side. Bliss's figures for Drosophila give a very similar result. The fact that the same temperature characteristic holds good for various different periods during one continuous developmental process obviously suggests that the shape of the curve must be much the same at different temperatures. The developmental process as a whole is therefore not deformed by change of temperature, but simply lengthened or shortened, its shape, as it were, remaining the same, or, in other words, either only one velocity constant is involved, or else, if more than one, velocity constants of the same temperature characteristic. To what extent this is really the case will appear presently. The only work on the temperature characteristic of the development of the chick is that of Brody & Henderson, who studied the effect of incubation at different temperatures upon the growth-rate of the embryo. Fig. 83 a taken from their paper shows how at any given day during development the weight attained is greater the higher the temperature; thus at the 17th day the average weight at 35-0° C. is 5-510 gm., at 37-3° 12-685 gm. and at 40-5° 20-427 gm. The upper part of the same graph shows the log. weight plotted against the age, as is usual in Brody's method, and the slopes of the resulting series of straight lines give the instantaneous percentage growth-rates, as follows : Phase or cycle I St 2nd 3rd °c. {k) {k) {k) 40-5 70 3^ 20 37-3 56 36 24 35-0 62 34 20 34-4 46 22 — 522 ON INCREASE IN SIZE [PT. Ill The instantaneous percentage growth-rate is shown in Fig. 83 b related to the temperature. The effect of temperature on the earliest ^ 3ms. 30 20 10 >>f., o105°F=40-5°C A 99°F=37-22°C A 95°F = 35°C X 94° F = 34'44°C • Temperature un ^^ y-:% ~^. \ known / ( "C y\ fl 90 f-^ zy y > >y ^ 8 A ' ^ if i, y y / / y i / / / y 5 / / / J / o?>^ .*. / fy V ^ X 3 . 2 /n-V /a / Xy y r? ^ i -y ^ i 1^ f '\^y / y V • Gms. / ) J K Y / ^ / X y y r 35 30 '/ y / 5 1 J^S>/

^^ 1 I 1 f / 4 i ti^/z •8 ^ ' J 1 7 t ' k ^ / / / / T/l 25 /^Li^jL^ / / y / •A / / \( yLif'7 <3'/_ ' / y A W • ^ i y y 20 -^ •3 vf ^A / y r J r / f'^ / 1 \ / -5? 15^ ( •2 ,/ /I y I ( w b /^ Y y V k >^ / 1 10 •1 > ^ l\ p / .'O H r ^■^ ^ y y < r-1 -^ ^ i H

-< ^ --^

D£ ^s 5 i 3 1 1. 1 .ba Fi 2 bio 1 a. 4 9^ 1 6 1 8 2 phase of growth is obviously the greatest of the three ; in other words, in the earlier stages rise in temperature has a proportionallygreater effect on the instantaneous percentage growth-rate than in the later stages. In the intermediate stage the rise is not so steep SECT, 2] AND WEIGHT 523 adalb bodytemp. 100K 70 65 60 j= 55 50 45 40 35 i. 30 25 20 15 with rise of temperature, and in the last stage, i.e. from about the 14th day onwards, the effect is nil, for the line is horizontal; or, in other words, the heat-regulating mechanism in the embryo is sufficiently developed at this stage to enable it to keep its body-temperature constant within the limits of the experimental temperatures so that they do not get a chance of affecting its growth-processes. It could indeed be argued that the three curves show a progressively increasing command of bodytemperature, so that higher temperature coefficients and steeper growth-rate/age curves are inevitably found the younger the embryo is. This question of temperature-regulation will be taken up again later in the book (see Section 4'2i). From these data the /x constants of Arrhenius' equation was calculated. For phase i it was 25,700 below 37° and 7100 above it, while for phase 2 it was 39,050 below 37° and 6500 above it. These values agree well with those obtained for other instances of embryonic growth, a fact which is especially important in view of the restriction of the other data to amphibia, fishes and arthropods. For the third phase there was obviously no ^ v^^ 'by^ X, K "V / / / ^^t \J0. & "^2 / / / V 14 -21 dax^s X3 35^ 37= 39° Temperature (c) Fig. 83 b. 41° 41-2'= temperature characteristic. Thus, before the age of about 13 days and between the temperatures of 35 and 37° increase of temperature by ten degrees more than doubles but does not quite triple the rate of growth, while after the 14th day the growth of the embryo is practically unaffected by change of temperature owing to its acquired homoiothermicity. Taking the values which have been obtained for /x for growth and development, it is clear that they all cluster rather closely round 524 ON INCREASE IN SIZE [PT. Ill certain definite magnitudes. Fig. 84, constructed from the data of all the relevant investigations, shows a frequency polygon for the values of fx associated with growth. Crozier's polygon for all fj, values (286 investigations) is described by a continuous line, and demonstrates the well-known modes at 8000, 11,000 and 16,000, as well as the smaller ones at 20,000 and above. These include all biological reactions which have been studied at different temperatures. But when only such fj, values as have been obtained for growth are o CT»6 O 4 C — Crosiers frequency polygon of all jx's. M Embryonic growth. — Growth as a whole I.e instances of adult + emb. l\ l\ I 1 / 1 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 JU XIOOO Fig. 84. plotted, the picture is quite a different one, for the modes do not correspond at all with those of the wider series. For the growthprocess, they exist at about 10,000, from 17,000 to 21,000 and at 27,000. The first of them falls between two peaks of Crozier's polygon, and the second, though represented there, equals, in this narrower consideration, the great 16,000 oxidation mode. The growthfrequency polygon obviously has little to do with oxidations, which are here graphically seen not to be the limiting factors in the growthprocess. If now embryonic growth alone is considered, an even narrower situation is disclosed, for embryonic growth has a peak at 10,000, practically nothing at all between 13,000 and 19,000, and SECT. 2] AND WEIGHT 525 after that a sustained high series until 25,000 is reached, followed by a spurt at 27,000. In a word, the peaks of growth as a whole and embryonic growth are very much alike, namely at 10,000, 20,000 and 27,000, but the middle one of these is specially associated with the development of the embryo. Now there can be no doubt but that the velocity of growth (reciprocal of the time required to attain a given size or stage of development) can be expressed well in its relation to temperature by the Arrhenius equation. Unfortunately, there is difficulty in deciding what is the exact meaning of the frequency with which the value of 20,000 occurs. There seems to be no reason why this should be so usual a feature of the time/temperature relation in embryos, and the modes at 10,000 and 27,000 are equally mysterious. No doubt it may be expected that further work, both in the direction of precision of data and in the direction of explanation of the significance of these values, will clear the matter up. The identification of the reactions responsible would certainly be a desirable thing. At present there is only one suggestion as to the meaning of a /i of 20,000, namely, its association by Glaser with "mobilisation hydrolysis", i.e. the production of oxidisable substances by hydrolytic action such as the formation of glucose from glycogen. The mode at 10,000 might be ascribed to synthetic processes, but this would apply better to the growth-values of 7000 or thereabouts. "I am quite unable to attach possible specific significance to temperature characteristics 8000 and 27,000", says Crozier in a private communication. "It will be realised that in the case of the earlier statements one is dealing with somewhat uncertain suggestions of which the chief value may be that they serve to suggest specific experiments. I should personally tend to deplore theories of the physico-chemical control of diflferentiation and development which might very easily be erected on the basis of such speculation." In these circumstances all that can be done is to note the values of fi which have been so far obtained for embryonic growth without attempting to give them any meaning. Perhaps the most important result of this kind of work is that no embryonic growth-process so far investigated turns out to have a temperature characteristic of 16,000, which is the typical one for oxidations. In this connection the paper of Loeb & Wasteneys in 191 1 is interesting; it was entitled: "Are oxidation processes the independent variables in life phenomena?" It occurred to them to 526 ON INCREASE IN SIZE [PT. Ul Table 74, Results of investigations of temperature characteristics. Species Mealworm {Tenebrio molitor) (pupa) 5J 5) 55 Fruit-fly {Drosophila) (egg + larva + pupa) 55 55 55 55 55 (imago) (egg+larva) 5J 55 55 (pupa) 55 55 55 Sea-urchin {Arbacia) (segmentation) 55 55 _ ••• Frog (Rana) (segmentation) Frog (Rana) (ist segmentation to disappearance of yolk-plug) Frog (Rana) (3rd segmentation to disappearance of yolk-plug) Frog (Rana) (later) ,, (till hatching) (early) „ (later) Cod {Gadus morrhua and aeglifinus) ,, {Gadus merlangus) Vlaice {Pleuronectes platessa) Japanese teleost {Hypomesus olidus) „ ,, {Plecoglossus altivelis) Water-beetle {Dytiscus marginalis) „ {Dytiscus semisulcatus) Water-flea {Cyclops) 55 55 Fruit-fly {Drosophila melamgaster) Water-flea {Pseudosida bidentata) ,, {Moina macrocopa) ,, {Simocephalus semisulcatus) 55 55 Hen {Gallus dormsticus) ... 55 55 55 55 ••• ••• 55 55 ••• ••• •" Fruit-fly {Drosophila melamgaster) 95 55 55 _ 55 Trout {Salmo/ario) „ J, Unless otherwise stated the lower /x is above 15°. compare the temperature coefficient of embryonic development in the sea-urchin's egg with the temperature coefficient of respiratory rate in the same material. Their figures for developmental rate gave Embryonic Growth growth Investigator 27,000 — Krogh 10,000 — ,, 27,000 27,000 Krafka 10,000 10,000 55 I9>920 — Loeb & Northrop 27,000 27,000 55 55 10,000 10,000 55 55 27,000 — 55 55 10,000 — ,, ,, — 41,000 Loeb & Wasteneys — 2 1 ,000 Loeb & Chamberlain — 12,400 Krogh — 21,900 95 — 10,800 95 — 19,100 Lillie & Knowlton 22,000 59 99 — 20,300 Krogh — 20,000 Barthelemy & Bonnet — 17,000 Hertwig — 24,000 Lillie & Knowlton — 11,800 20,000 Dannevig — 20,000 Johansen & Krogh — 23,700 Higurashi & Tauti — 23,000 Higurashi & Nakai — 19,300 Blunk — 20,000 _99 — 10,400 Ziegelmayer — 155700 99 335210 Bliss 16,850 — 99 7,100 — 99 19,800 — Brown 28,500 — ,, 17,210 — 99 7,410 — 99 16,950 — ,, 4,780 — 99 25,700 Brody & Henderson — 7,100 99 99 — 395O5O 55 99 — 6,500 99 99 27,800 — Plunkett 17,100 — 99 9,000 — 99 — 24,500 Gray — 20,000 Maitland SECT. 2] AND WEIGHT 527 very considerable slowing at the low temperatures, so that although the coefficient between 17 and 27° was 2-6, between 8 and 18° it was 6-0. The temperature coefficient for the respiratory rate, however, did not show this, and, while it was 1-5 approximately round 30°, never rose higher than 2-45 even including the data for between 3 and 13°. There was thus no correspondence between the two sets of coefficients below 15°. This circumstance, they thought, harmonised with the fact that potassium cyanide in concentrations which stopped development altogether still permitted 25 per cent, of the oxidations to proceed, and further, with the fact that when no development was going on normally, i.e. before fertilisation, there was still a certain respiration. Of the two processes the morphological development rate seemed to be the more easily abolished by external agents. But above 32° when cleavage became impossible, respiration was also impossible, and the oxygen uptake rapidly fell off. "These facts", said Loeb & Wasteneys, "do not contradict the view that oxidations are the independent variables in developmental processes. But on the other hand they do not furnish any convincing evidence in favour of such a view." The whole incident, as a matter of fact, may now be regarded as an instance of the incapacity of the ordinary van't Hoff equation to differentiate between two alternatives in a complicated biological system, and the question which Loeb & Wasteneys were trying to answer may be said to be solved by the frequency polygon of temperature characteristics. The results of all this work may be summed up by saying that by the use of the heat effect on embryonic growth and development a kind of dissection has been made of the process, for at one temperature one controlling reaction or pacemaker is laid bare, as it were, and, at another temperature, another controlling reaction. This exposing method has led to the unveiling of four or five reactions of quite clear-cut temperature characteristics, which according to the temperature and the circumstances may be in control of embryonic growth. What these reactions are we do not yet know, but admittedly a good step forward has been taken by the identification of the heats of activation of their active molecules. Crozier has discussed in some detail the relation between the temperature characteristic of growth and the autocatakinetic growth curve. Difficulties arose out of the experimental fact that temporary alteration of the temperature during a period of development 528 ON INCREASE IN SIZE [pt. m modified the temperature characteristic for the remainder of development. This it could not be expected to do on the Robertsonian view, for where there is only one velocity constant, as in his presentation, its temperature characteristic must always be the same, and changing the temperature would only multiply the time coordinate by some constant, and drag out or compress the time taken to complete the autocatalytic curve. Within one cycle the temperature characteristics should be the same for all partial developmental periods as well as for the whole. This, however, was found not to be the case by Bliss for Drosophila, by Brown for the cladoceran Pseudosida bidentata, and by Titschak for the clothes-moth Tineola biselliela. If an animal is allowed to develop for 50 per cent, of the total normal time at 15° C, and is then transferred to a temperature of 25° C, it may take less or more time to finish its development than would be predicted on the basis of the fact that it has still 50 per cent, of its normal time to go at 15°. Brown expressed the differences as per cent, gain or loss in time under such treatment, and found that they were, though small, quite significant statistically. This means that the growth equation of Robertson, , = Kx (a — x) dt (where a is the initial amount of growth-forming substance, x the amount formed after time t and K a constant), cannot hold, and must give place to an equation with two velocity constants. Crozier has suggested one, the differential form of which is ~={K, + K,x) [A - x), where Kj^ is the velocity constant proper to the reaction A^x, but in the absence of the catalytic effect of x, while K^ is the velocity constant associated with the process when x is functioning as a catalyst. The point of inflection will then be The integral form of the differential equation is / = , A {K^x + K^) K^ + K^A = K^{A- x) SECT. 2] AND WEIGHT 529 and though the form of its curve is sigmoid the point of inflection depends on the ratio of the two velocity constants, so that it may be very asymmetrical. Before embarking on the exposition of the applications of the Arrhenius formula to growth in general and embryonic growth in particular, it was said that there was much difference of opinion as to whether the time/temperature relation could be expressed best by a hyperbola or an exponential curve. It was often pointed out by Loeb and others that even a straight line might be produced by the action of various factors on a true exponential curve causing it to flatten out. Snyder suggested that an important factor which might be expected to have such an effect was the protoplasmic viscosity. This idea has since had a considerable popularity, although Crozier has vigorously combated it. "The attempt to introduce considerations of protoplasmic fluidity, presumably as influencing diffusion, requires a theory of the general control of organic activities by the whole body of the cell rather than at surfaces. This is unnecessary", says Crozier, "and at present inadmissible." But it is by no means easy to see why, and the possibilities contained in corrections for viscosity should certainly be explored. The difficulty is that viscosity varies in different ways with temperature according to the animal or plant used. Some authors, moreover, seem to wish to obtain a constant Q_io, although even in inorganic in vitro reactions Q^io is never constant, but increases as the temperature is lowered. Pantin, for instance, corrected various temperature coefficients, on the basis of the viscosity changes in Nereis eggs, and found a much greater approach to constancy, though even after correction, the Q^^q rose a good deal at the lower temperatures. Krafka's work, already referred to, is rather an obstacle to this point of view. The same line of thought was carried further still by Belehradek, who criticised the use of the Arrhenius equation on the ground that, like the van't Hoff" formula, it also did not give a constant for all temperatures at which normal development will proceed, but shows up critical points which may or may not be real. This criticism was also made by Heilbrunn. Belehradek used a modification of the Esson-Harcourt equation, which he found would represent the time/ temperature relation fairly accurately: y = ^ or logy = log a — b log x, NEI 34 530 ON INCREASE IN SIZE [PT. Ill where jy is the time necessary to accompHsh a given amount of e.g. embryonic development at the temperature x and a and b are constants. Thus log. time/log. temperature gives a straight line. The constant a varies according to the unit of time used, but the constant b is independent of the time-unit and the actual velocity for it represents acceleration. If ^ = i then y = ajx, which is the equation for a rectangular hyperbola, and covers the work of Peairs on insect development as a special case of the general law. Again, the straightline relation of Krogh and others is another special case, for, according to it, V ^ kx where v is the velocity and x the temperature and, as jv = ijv and k = i/a, we get y = a/x, which is the original equation once more. As b is usually greater than i, Q^^o usually declines with rise of temperature. Belehradek regards the constant b as a true temperature coefficient, for it does not change with temperature, and he has calculated it for various systems which are relevant here. Thus he gets for the embryonic development of Copepod Cyclops fuscus (Ziegelmayer) ... Fruit-fly Drosophila melanogaster (Loeb & Northrop) ... Frog Rana virescens (Lillie & Knowlton) Hen Callus domesticus (Fere) and for the early segmentation of Sea-urchin Strongylocentrotus lividus (Loeb) Roundworm Ascaris megalocephalus (Faure-Fremiet) It would appear that b increases with age, for Fruit-fly Drosophila melanogaster (Loeb & Northrop) : Larva Pupa b i-i6 2-IO 2-36 4-10 I St 0-99 2-50 2nd I-I3 and and and Water-beetle Dytiscus semisulcatus (Blunk) : Embryo Larva (ist instar) ... Larva (2nd instar) Larva (3rd instar) Prepupa Pupa Frog Ranafusca (Krogh) : FertiHsation to ist cleavage End of 1st cleavage to closure of neural fold Closure of neural fold to appearance of external gills Appearance of external gills to 3-branched gills 3-branched gills to 7 mm. embryo ... • ... 7 mm. embryo to 7-8 mm. embryo Typographic beetle Tomicus typographus (Hennings) : Embryo Larva 210 2-28 i-io 1-14 1-26 1-38 1-48 I -60 I -20 1-64 1-76 1-92 1-69 2-52 2 -02 3-52 SECT. 2] AND WEIGHT 531 B^lehradek interprets b as being a measure of protoplasmic viscosity, so tliat in his opinion all temperature effects on growth can be regarded as primarily effects upon viscosity, and only acting indirectly upon the rapidity of the growth-process. Whether this point of view will prove to be either more fruitful or more correct than that of Crozier and his collaborators cannot at present be decided. Crozier & Stier have, however, shown that the fit of B^lehradek's formula in one case, at any rate, is not at all good, while Belehradek has strengthened his case by studying the time factor in cooling and heating protoplasm. But it is interesting that the constant b should increase with age, for protoplasmic viscosity almost certainly does, and this has an obvious importance in view of the gradual loading up Exponential curve Catenary Fig. 85. Hyperbola of the cells of the body with paraplasmatic substances. Belehradek has answered Crozier's criticisms and the papers must be consulted for the details of the argument. Janisch represents the time/temperature relation not by a simple exponential curve, nor by a hyperbola, but by a " Kettenlinie " or catenary curve. The reciprocal of this is not a straight line or a sigmoid curve rising most rapidly at the point of inflection; it is a sigmoid curve rising most slowly at the point of inflection. Such a complex exponential curve should therefore be produced when the velocity/temperature graph is drawn, and Janisch actually showed that Krogh's straight lines turn into curves of this character when the points which Krogh neglected at the upper and lower ends are taken into consideration. Just the same can be shown for the figures of Sanderson and Peairs. Fig. 85, which is taken from Janisch, shows the relations between these curves. Fig. 86, also taken from Janisch, shows a replotting of the data of Sanderson for Margaropus annulatus. The time of development is shortest at 28°, and, just as Faure 34-2 532 ON INCREASE IN SIZE [PT. Ill Fremiet showed for the segmentation of Ascaris, it lengthens a Httle on the high-temperature side of that. The points are well described by one of Janisch's catenary curves. The velocity of development, i.e. the reciprocal of the development time, correspondingly rises to a maximum at 28°, and has a sigmoid form on each side of this maximum. This peaked curve, part of which, it should be remembered, corresponds to the Krogh "straight-line", is identified by Janisch with the peaked curve which Duclaux in 1899 pointed out would be produced by the operation of two separate factors giving complex exponential curves, i.e. differentially affected by temperature change. This peaked curve could be either symmetrical or asymmetrical. 1 _ • / , » /'» ^, /• X \ • ^""^-.ui^ . iu^ .^ ^^^^ 1 1 1 1 1 1 1 1 1 1 1 1 -^ 160 140 120 100 I 80 Cl 60 40 20 38° 36 34 32 30 28 28 24 22 20 13 16 14 12 10 8° Temperature Fig. 86. Perhaps further work along these lines will lead to conclusions about critical temperatures and controlling reactions which would support those of Crozier. Janisch gave a full mathematical treatment of these questions, and also reported an investigation of the effect of heat on the embryonic development of the "Mediterranean flour-moth" Ephestia kuhniella (see also Hase). The time/temperature relation here also can be expressed by an equation of the form y 2 % where m and a are constants and y is the time taken to complete a given amount of development at time x. Such presentations SECT. 2] AND WEIGHT 533 of the data are very interesting, and possess advantages as well as disadvantages in comparison with the equations of definite physical meaning, such as that of Arrhenius. For, although they do not seem to give immediately so great an insight into the fundamental processes of the living cell, they are in value independent of advances in physics which may destroy the basis on which the other kind of treatment rests. In other words, their status as short descriptions of the phenomena is safer if not so exalted. I have now completed what it was necessary to say about the effects of temperature on embryonic growth and development. This knowledge is very valuable, but the unveiling of the limiting factor by temperature control, and the identification of the "master reaction" by fitting growth-curves, are perfect examples of what may be called the "short-cut" method. To a certain type of mind the attractions of the short-cut are too great to be resisted, but, as we have seen, the results are liable to be enigmatic. 2-17. The Effect of Light on Embryonic Growth The influence of light upon the course of embryonic growth has been little studied, though the results of such work cannot now be looked on with the scepticism of twenty years ago. Schnetzler's paper of 1874 was the first serious contribution to the study, although Edwards fifty years before had stated that frog eggs would not develop at all in the dark, a groundless assertion which had been disproved by Higginbottom and McDonnell. Schnetzler; and Baudrimont & M. de St Ange also found no difference in hatching time, though the former thought that the growth of the tadpoles was more rapid in the light than in the dark. This was substantiated quantitatively by Auerbach and by Yung, who placed freshly laid frog's eggs in vessels containing 60 eggs in 4 litres of water, some in diffuse light, others in the dark. Yung's figures were as follows: Days Size of light tadpoles compared with dark (100) development Length Breadth 25 125 124 30 117 118 60 106 107 It is interesting that the light effects diminish proportionally with 534 ON INCREASE IN SIZE [pt. iii increasing age. Lessona and Camerano brought forward field evidence which agreed with the controlled results of Yung, who was later confirmed by Perna and by Chiarugi & Livini. Yung also found that the eggs of Salmo trutta hatched i day earlier when kept in the light than when kept in the dark, and that the eggs of Limnaea stagnalis, the pond snail, hatched in 27 days in the light but 33 days in the dark. Davenport, following earlier speculations of Millet and Blanc, has commented upon the regularity with which developing embryos are hidden away from the light either in the egg or in the uterus, but as this is not the case with animals such as the echinodermata there seems no great significance in the idea. Bodine and Carothers have observed a marked inhibitory effect on embryonic development of certain orthoptera [Melanoplus differentialis, Chortophaga viridifasciata, Circotettix verniculatus) of direct sunlight (not the heat accompanying it) and Miyakawa affirms the same for Schistosomumjaponicum*; but, on the other hand. Page, taking like precautions, reported that shad eggs could be made to hatch 12-15 hours before the controls by allowing direct sunlight to fall upon them. Ruffini's curious observation that the eggs oiBufo vulgaris in their jellies always orient themselves with their animal poles pointing toward the light has never been confirmed, but may have significance. According to Goodrich & Scott light has no effect on chick embryo cells growing in vitro. Other researches have been made with the object of identifying the most active wave-length. Beclard very early placed the eggs of Musca carnaria behind screens of different colours, and found that the embryos in the green light were the least developed after a definite period, the rest following in the order, yellow, red, blue and violet, in which last light they developed most rapidly. The retarding effect of the green was also observed by Schnetzler. Davidson's results are untrustworthy. Very striking effects were observed by Yung, whose papers are the most often quoted : Length of tadpoles Colour of light r I month 2 months White 100 100 Violet 117 134 Blue 105 107 Yellow 99 102 Red 83 86 Green 77 All dead

See also the work of Frederich & Steiner. SECT. 2] AND WEIGHT 535 Plus Minus 2-5 — — 1-3 — 45 — 4-8 — 6-9 — 8-9 As regards echinoderm embryos, we have the following table of Vernon : Length °/^ change from white length Semi-darkness Complete darkness Blue Green Red Yellow Driesch stated that these colours had no effect on Echinus embryos. Here the favourable action of violet or blue and the unfavourable action of green are not apparent. Nor in the case of trout fry could Schondorff find any differences between the colours. The hatching time of various eggs, however, according to Yung, does bring it out again, as Table 75 shows, and the qualitative experiments of Schenk on Rana and Bufo eggs and Fatigati on other material (infusoria) demonstrate the same relation. Yung's figures: Table 75. Hatching time in days Violet Blue Loligo vulgaris Salmo trutta ... Limnaea stagnalis 50 32 17 53 35 19 Yellow 58 34 25 Red 58 36 Green Dead 36 Dead White Dead 35 27 In 1 91 3 Grein subjected the eggs of Gadus virens to light which had passed through the following screens : Wave-length allowed to pass A Red ... 610-710 /x/Li B Red ... 600-720 C Red ... 600-700 D Green ... 440-460 E Blue ... All save light green and dark yellow Hatching times were quicker under D than under A, B or C, and quicker under A, B or C than under E. Supino's results, expressed in percentage of embryos hatched after a fixed period, were as follows : Blue ... 65 Yellow ... ... 43 White ... - 53 Red -. 39 Darkness - 45 It may, on the whole, be concluded that the wave-lengths contained in diffuse daylight which affect embryonic development are 536 ON INCREASE IN SIZE [pt. iii the short ones at the violet end of the spectrum. It would be verydesirable to go into the whole subject anew in the light of modern technique and modern conceptions. 2- 1 8. The Effect of X-rays and Electricity on Embryonic Growth Other forms of radiant energy have also been investigated, but the position is very complicated and unsatisfactory. Oilman & Baetjer in 1904, judging from morphological and cytological evidence, concluded that in Amblystoma, X-rays at first accelerated development and then retarded it so that eventually the exposed embryos were smaller than the controls. For the silkworm Hastings, Beckton & Wedd have asserted that X-rays accelerate the developmental time, and (Hastings) that the secondary radiation excited by irradiating copper retards the development of the silkworm. Ancel & Vintemberger in an elaborate and long research, decided that in the case of the chick and the frog, X-radiation had neither an accelerating nor a retarding influence. Nevertheless Colwell, Gladstone & Wakeley believed that they had evidence of retardation by X-rays in the chick, as the following table shows: Table 76. Length of embryo Thickness of on the eighth Daily dose (P.D.) filter (mm.) day (mm.) (control) — 22 I 4 4 I I 4 I 2 4 f i 8 t I 10 1 2 12 i I 13 i 2 13 i i 15 i I 15 i 2 15 Their second paper, though mainly morphological, confirmed the earlier one. Richards obtained acceleration working with the egg of the mollusc Planorbis, and the opisthobranch Haminea virescens. Packard working with Arbacia, Lazarus-Barlow & Beckton with Ascaris, Bohn with Strongylocentrotus eggs, found that short exposures to radium accelerated development and long ones retarded it. Lazarus-Barlow & Beckton found that /S and 7 radiations alone had SECT. 2] AND WEIGHT 537 the same effect as all three together though the a radiation was 100 times as much as the other two. Forsterling obtained retardation of growth in rabbit embryos by irradiating the mother, and other work on mammals was done by Cohn, by Lengfellner, and by Bagg, but the conditions are there so complicated that it is not worth discussing. The effect of electricity upon embryonic growth has been fairly often tried, but usually in an unintelligent way. Eggs have simply been placed between the electrodes or between the poles of a magnet, and conclusions have been drawn which would probably not bear statistical examination for a moment. Thus Rusconi in 1840 stated that frog's eggs hatched more quickly under the influence of electric currents than otherwise. Lombardini remarked the same effect in amphibian development but noted a large number of abnormalities, as did Fasola. Windle, working with eggs of the trout, found no acceleration of development either under the influence of electric currents or of a large magnet, but he noticed that the hatched trout died very quickly. Rossi found many anomalies in the eggs of urodele amphibia subjected to electric currents, but no other effect. Slater did not find any result at all when he subjected silkworm eggs to a strong magnetic field, nor did Maggiorani when he subjected hen's eggs to one. Finally, Benedicenti found no effect of constant weak currents on the development of echinoderm eggs. The more recent work includes that of Scheminzki on the trout — he could observe no acceleration of development when the eggs were subjected to constant sub-lethal currents, but the membrane at the end of the development was weaker than usual so that the embryos tended to hatch early. Gianferrari & Pugno-Vanoni passed currents of 9000 volts at 450,000 periods/second through suspensions of trout {Salmo lacustris) and echinoderm {Echinus esculentus) eggs, but the effects observed were merely teratogenic, and the abnormalities produced did not differ, apparently, from those produced in other ways. The subject still awaits an investigator who will sweep up all this debris into some coherent theory of the action of electrical currents and magnetic fields on embryonic development. The other factors, such as osmotic pressure, which have been shown to affect growth so greatly in the post-natal stages and in plants, do not exercise so much influence on the foetus, guarded as it usually is from the world around it by the egg-shell or the uterus. Their effects have therefore not been much studied in the case of 538 ON INCREASE IN SIZE [pt. iii embryonic growth, and the few data that do exist on the subject will be presented in the section on comparative susceptibility of the embryo at the different stages in its development. The section on biophysical phenomena also contains information on cognate points. The process of cell-division, as such, is of course influenced by a great variety of factors and substances, as the following table shows : Table 77. Substances which accelerate cell-division Authority Heat ... ... ... ... ... Laughlin X-rays ... ... ... ... ... Gilman & Baetjer Radium emanation ... ... ... Packard; Bohn; Shumway; and Buddington & Harvey Th>Toiodin ... ... ... ... Richards Adrenalin ... ... ... ... Chambers Alcohol Calkins; Woodruff Potassium hydrogen phosphate ... Woodruff Potassium sulphate ... ... ... „ Potassium bromide ... ... ... ,, Oxygen ... ... ... ... Godlevski Sodium and potassium hydroxide . . . Richards ; Loeb Pilocarpine ... ... ... ... Richards Ammonium hydroxide ... ... ,, In most cases, however, the effects produced by these agents are complicated, and the original papers should be referred to. Naturally, all these substances and factors exercise a depressant action on celldivision if they are applied in too brutal a manner. Moreover, though they tell us a certain amount about the nature of mitosis, they do not much assist in the understanding of the metazoal growth-process, organising mitoses as it does on the large scale. As Richards says, " The advantage gained in the segmentation-stages may later manifest itself in more vigorous larvae than in more rapidly developing ones ", so that agents which accelerate cleavage may not accelerate hatching-time. The study of these agents has not led so far to any great advance in our knowledge of the essential nature of growth and development. 2-19. The Effect of Hormones on Embryonic Growth As regards the effects of endocrine organs upon embryonic growth, practically nothing is known. Willier, who made chorio-allantoic grafts of thyroid in chick eggs, observed that the host embryo was always smaller, and in some cases as much as one-third smaller, than the control. Hanan later, after a tliorough study of the difficult question of appropriate dosage, injected 1/600 of a mgm. of thyroxin SECT. 2] AND WEIGHT 539 into the air-space of chick eggs on the 5th day of development. The effect was absolutely nil, as is seen from Figs, 87 & 88 where his points are plotted both for wet and dry weight. There is thus a certain contradiction between the results of Willier and those of Hanan ; Okada afterwards repeated Hanan's experiment, and observed no effect on the general appearance, but an 18 % decrease in length and weight of the chicks from the injected eggs. There were characteristic histological changes, mostly in the direction of precocity. Okada's results would fit in with Butler's finding of a decrease in cell-division rate under the influence of thyroxin in Arbacia eggs. Pituitrin has also been injected into hen's eggs, by Cunningham & Stanfield, who obtained the following results : Average weights (gm.) Injection 7th day, opened gth Injection 7th day, allowed to hatch Injection 6th day, opened 8th /■ > Control Injected embryos embryos 1-578 i-goi 37-4 400 0950 I 203 540 ON INCREASE IN SIZE AND WEIGHT [pt. iii This may mean that growth is accelerated, but a larger number of experiments with more complete statistical treatment will be required to settle the question. A word should be said about the possible total reversibility of growth. Regression to the embryonic state was first brought about by Caullery, and recently Davydov has succeeded in reducing the nemertine Lineus lacteus to a state which cannot be distinguished from its blastula. Nothing is known of the chemistry of this process. In conclusion it may be said that modern embryology has more and more come to follow the lead of Leonardo da Vinci in subjecting all the aspects of the growth of the embryo to exact measurement. In doing so, it has, as we have already seen, put itself in a position to fuse its studies with those of biochemistry and biophysics. The fruit of this union can only be the understanding of the molecular mechanisms underlying embryonic growth and development, from the egg-cell into the newly hatched animal, whose shape and constitution must in the last analysis be regarded as among the properties of what we call matter.

Section 3 On Increase in Complexity and Organisation

3-1. The Independence of Growth and Differentiation

The previous I has been concerned solely with the question of increase in size and in mass as the embryo develops; other synchronous processes were perforce neglected. These may be comprehended in one word, differentiation, or increase in complexity and organisation both macroscopic and microscopic. "All generation", said Sir Kenelm Digby, "is made of a fitting, but remote, homogeneall compounded substance; upon which outward Agents, working in the due course of nature, do change it into another substance, quite different from the first, and do make it less homogeneall than the first was. And other circumstances and agents do change this second into a third, that third, into a fourth ; and so onwards, by successive mutations that still make every new thing become lesse homogeneall than the former was, according to the nature of heat, mingling more and more different bodies together, untill that substance be produced, which we consider to be the period of all these mutations." Or, as William Harvey puts it, "For though the Head of the Chicken, and the rest of its Trunck, or corporature (being first of a similar constitution) do resemble a Mucus, or a soft glewey substance: out of which afterwards all the parts are framed in their order; yet by the same operation, and the same Operatour, they are together made and augmented: and as the substance resembUng glew doth grow, so are the parts distinguished. They are Generated, Altered, and Formed at once". Having discussed then in the preceding section the "augmentation" of the embryo, we have now to discuss its "making" or in Harvey's other term, its "framing". Framing is undoubtedly not an increase in complexity alone. Woodger has well said, "It is often stated that organisms are just complicated physico-chemical reactions, and it is because they are so complicated that biology has so far made so little progress. But it is evidently not simply a question of complicatedness, because there are plenty of compUcated goings on in the world which no one 542 ON INCREASE IN COMPLEXITY [PT. Ill would mistake for organisms. The 'something going on' which we call a thunderstorm, for example, is very complicated. An organism from some points of view is comparatively simple, otherwise biology would not have got as far as it has, and this simplicity appears to be the outcome of its organisation. What happens in development is not merely an increase in complexity nor an increase in spatial structure, but a gradual rise in the level of organisation of the developing organism". It has often been said that the interesting thing about any magnitude is not so much its absolute as its relative size, and not so much its relative size as the rate at which its relative size is changing. The first question to ask therefore in the discussion of the physicochemical aspect of differentiation is whether the rate of differentiation is the same as the rate of growth at each embryonic stage. That the two processes are independent in the sense that one can be induced without the other has long been known. Panum, for instance, as long ago as i860 observed chick embryos with well-developed blastoderms but no primitive streak, but he did not associate this with any definite causative factor. Then Broca found he could get this growth without differentiation by keeping fertile eggs for a month or more at room temperature before incubating them, and Dareste observed that very high as well as very low temperatures maintained during the first 24 hours of development would bring about the same effect. Rabaud confirmed this : for further information see the review of Tur. Finally, Edwards found that cell-division without embryonic organisation always occurred in eggs incubated from 20 to 27°. Continued growth may take place also in the primitive streak stage as well as in the simple blastoderm when there is not a sufficient degree of temperature to permit of normal differentiation. Fig. 89, taken from Edwards' paper, shows the effects he obtained. The controlled action of temperature can also bring about nuclear division without corresponding cell-division in Echinus^ according to " 12° 23° TemperaJture

Fig. 89.

Driesch, and many other workers have observed these polynuclear undifferentiated masses in other embryos with abnormal temperatures. Even in bacteria growth in size can take place without cleavage, according to Henrici, and the work on the endocrine control of amphibian metamorphosis affords other instances of continuous growth without differentiation. The most remarkable case of this, perhaps, is contained in the work of Hoadley, who transplanted embryonic organs of the chick on to the chorio-allantoic membrane and allowed them to develop there until they had reached a degree of differentiation equivalent to that of the controls. Then by making and weighing wax models he ascertained the relative weights, and always found that the controls were much heavier. But there was a direct relation with age, for the younger the transplant at the time of transplantation the smaller the eventual (fully-differentiated) organ, thus: At the time of transplantation A Control : Transplant = r ■ ■\ Embryo age No. of X Organ (hours) somites (after about 8 days) Spinal cord 48 28 0-353 Eye 48 28 0-219 >> 35 14 0-098 }} 20 0-0136 „ 40 0-00075 Another instance of the independence of growth-rate and differentiation-rate is afforded by the genetic races of rabbits which differ considerably in size. At birth the large-sized race is twice as heavy as the small-sized race, and at the adult stage, three times as heavy. In 1928 Painter observed that equally differentiated 12-day embryos showed the characteristic weight-differences and thought that the effect might be of endocrine origin, but Castle & Gregory were later able to trace it back as far as the morula stage. The large-sized race showed consistently more rapid cell-multiplication and size increase in the embryonic period with unaltered differentiation-rate. There is no difference in egg-size: but the large-sized race grows more ^ ^ ' ' Embryo age Average no. of Diameter of in hours blastomeres blastodermic vesicle Large-sized race ... ... 48 21*75 — Small-sized race ... ... 48 14-00 — Large-sized race ... ... 100 — 31*4 Small-sized race ... ... 100 — 16-3 Another good instance of this dissociation of the fundamental embryogenic processes is seen in the interesting work of Twitty, who studied the development and the nature of the action of the ciha on the skin of amphibian embryos, previously described by Assheton and Woerdemann. He found that the polarity of the ciliary cells is determined during the closure of the neural folds, for cilia grafts rotated 1 80° before that stage beat in the same direction as the cilia of the adjacent ectoderm, but cilia grafts transplanted later retained their original direction of waving. Thus the determination here (in Amblystoma punctatum) occurred much later than the main point of chemodifferentiation (see p. 575). Now in embryos allowed to develop at low temperatures, Twitty found that this ciliary polarity appeared at a much earlier stage. Evidently the determinative process had been thrown out of gear with the morphological ones by the cold, and the two were proceeding more or less independently. Again, certain treatments, as is well known, inhibit segmentation, and cause the production of cilia, while histogenesis and organogenesis are easily separable as is seen in teratological experiments (Ranzi) and in explantation work (Hoadley)*. 3*2. Differentiation-rate Under certain conditions then, such fundamental processes as growth and differentiation can be shown to be proceeding independently of each other. That their normal velocities differ might or might not be the case. All who have had any practical contact with embryology know that there is much more difference in shape and form between a chick embryo of the 2nd day and one of the 5th than there is between one of the 12th day and one of the 15th. But the difficulty has always been to evolve some method of measuring change in shape — a more elusive entity than change in weight. Murray, however, has made an admirable attempt in this direction, choosing as the organ for investigation the heart, which, as we have already seen in the work of Schmalhausen, appears to keep pace exactly as regards growth with the embryo as a whole. He tested this by measurements of the surface area of the organ, calculating the percentage growth-rate of the heart from Cohn's work with a projectoscope and a planimeter. The result was one of the usual descending curves which closely resembled that for the embryo as a whole, and the log. surface area of the heart/log. age of embryo relation was also a straight line.

Thus Waddington, working with chick embryos cultivated in vitro, found that the i2-somite stage was much smaller than the corresponding stage in the egg. bo 35 546 ON INCREASE IN COMPLEXITY [PT. Ill "No satisfactory method", Murray said, "of measuring changes in form quantitatively is known, so that it was necessary to resort to the expedient of selecting forms spaced by a visual impression so as to represent approximately equal degrees of gross change. In other words, from a series of drawings made at frequent intervals, certain ones were chosen which seemed by inspection to be equally spaced from one another in respect to their relative complexity of form. The test is thus necessarily arbitrary and open to criticism because of its <6 O O o 100 o-| 50 k o^ N '^ -— o Ds.ys z 6 8 10 Incubd^tion Fig- 91 a subjective nature. By taking the average result of many eggs it was then determined what were the exact incubation ages of the embryos with heart forms such as those selected." The illustration shows these clearly (Fig. 90). The reciprocals of the time intervals between successive drawings were used as rough criteria of the rate of form development. The graph showing this done is reproduced in Fig. 91, from which it is clear that the rate of evolution of external form falls precipitously at first, and then ever more slowly, essentially resembling in this way the instantaneous percentage growth-rate, according to Schmalhausen, or the averages of the steps in the fall of the instantaneous percentage growth-rate, according to Brody. Growth SECT. 3] AND ORGANISATION 547 and form would therefore seem to be what Murray calls "aligned" with one another, for the most marked form changes occur at the initiation of embryonic development when the growth-rate is at its highest. Murray's interim method affords a basis of quantitative comparison between growth-rate and differentiation-rate. But if these two processes change together with like velocity, others do not. The shifts of chemical and physical change within the embryonic body follow an entirely different course In 1925 I had drawn attention to the fact that, making allowance for the increase in size of the embryo and for its consequently increased daily turnover of matter and energy, chemical activity seemed to be more intense at the end of development than at the beginning. "One pictures", I had said, "the gradual elaboration of structure up to the beginning of the third week, followed then by a burst of chemical activity consequent upon the assumption of function by the organs already formed. In this hypothesis is involved the view that in the first fortnight chemical change is limited to those compounds which are required mainly for structural purposes, while toward the end of incubation the opening up of functional operations causes marked and profound chemical changes of other kinds." This was proposed in connection with the three periods which modern embryologists have come to recognise in all embryonic development. The same idea was elaborated further by Murray in relation to change in growth-rate and differentiation-rate. Taking the rate of change in chemical constitution, he concluded (largely from his own experiments) that the most marked changes occurred after the loth day, not before it, as is the case with growth-rate. "Internal integration", he said, "may be regarded as a process characterised by the concentration of solid substances within the body, whereas chemical differentiation is a change in the composition of the solid substances thus integrated." Fig. 92, taken from Murray, demonstrates these relationships, for it may be observed that the percentage of ash in the embryo, the percentage of total solids and the percentage of fat, together with the rate of production of carbon dioxide, all rise or fall in the same way, i.e. more rapidly at the end of development than at the beginning, and precisely opposite to the growth-rate and the differentiation- rate. Thus, just as morphological growth and differentiation take place more rapidly the younger the embryo, so chemical growth and differentiation take place more rapidly the 35-2 548 ON INCREASE IN COMPLEXITY [PT. Ill older the embryo. A note of caution must be inserted here, however, in view of the fact that, until we know more about the behaviour of individual organs in the light of these conceptions, we cannot estimate the part played by ossification, for instance, in producing these overall results. Murray also afterwards found that the curve Per Pep cent cent 9 11 13 Incubation age Fig. 92. relating age with oxygen absorbed by the embryo per gm, per day did not decline quite like the carbon dioxide evolution curve, i.e. slowly at first, and then quicker and quicker, but was rather S-shaped, the slowest rate occurring at the loth day. His line for this, however, depended for the initial fall exclusively upon one very high point for the 6th day, and this may quite possibly have been wrong, in which case the curve for oxygen consumption would fall with the curve for carbon dioxide output. SECT. 3] AND ORGANISATION 549 Murray next showed that the curves for growth-rate of heart tissue in explanted culture, obtained by himself working with Cohn & Rosenthal (see p. 461), fell in the same way as those for the intact embryo in its egg, but that the curve for latent period (see p. 462) behaved, on the contrary, exactly like those for integration and Deo/35 11 13 Incubation age Fig- 93 chemical differentiation, rising very slowly at first, and thereafter more rapidly. Thus chemical constitution, metabolic rate, and latent period of explant growth are to be correlated and distinguished from growth-rate both in the intact animal and in tissue culture. Fig. 93 shows the difference between growth-rate and metabolic rate, as judged from the rate of carbon dioxide production. As a result of this analysis of embryonic development, at least five 550 ON INCREASE IN COMPLEXITY [pt. iii fundamental processes may be distinguished, falling into two groups. The first group consists of those processes which have their maximum rate of change early in development, and it includes both growthrate and differentiation-rate. The second group consists of those processes which have their maximum rate of change late in development, and it includes, firstly, the concentration of solids in the embryo (called by Murray "chemical growth ") , secondly, the increase in complexity or change of composition in the solid substance (called by Murray "chemical differentiation") *, and thirdly, the metabolic rate or respiratory intensity. These two groups of processes correspond to two type curves which in some circumstances show skew symmetry round a central point. The first group has been termed by Murray the group of "primary integration" and the second one the group of "secondary integration", and he has in various papers associated the two groups with the primary and secondary redistributions of the evolutionary process in Herbert Spencer's scheme, or, as they are often called, simple and compound evolution. Murray found that the surface volume ratio was also a member of the first group, changing most rapidly in the earliest stages and falling in a way not unlike the growth-rate. He calculated it from the formula of Meeh : 2 S= KW\ and, although such an application to the chick embryo rests inevitably on several unproved assumptions, it is probable that his curve is not wholly misleading (see Fig. 94) . More interesting still, he found in a later paper that the absorption curve (the grams absorbed per gram dry weight per day) was not a member of either group, for it was sigmoid, falling rapidly for a time, then less rapidly, and then more rapidly again. His absorption curve, however, was derived from the oxygen measurements which have already been criticised, a fact which might account for its S-shaped character. About the same time, I also calculated the curve of absorption-rate, not from respiration experiments but from direct measurements of protein, fat and carbohydrate, and I obtained a curve belonging to group I, i.e. changing most rapidly at the beginning of development. It is very interesting that the curve of absorption-rate should be found to follow the curve for surface/volume ratio (see Fig. 253).

This process of "chemical differentiation" must not be confused with "chemodifferentiation " (seep. 571). SECT. 3] AND ORGANISATION 551 Murray's presentation of the facts does bring an order into a realm where order is much needed. But it is not ahogether easy to see why increase in dry weight should be regarded as "chemical growth", for water is just as much a chemical compound as anything else. Why should the title of chemical compound be restricted to those bodies which do not happen to be liquids at room temperature? And the wetness Deyas 11 13 Incubation age Fig. 94 of the embryonic body must also exercise a profound influence on the speed and nature of the chemical reactions going on within it, as is emphasised in the work of Ruzicka and his school. There is therefore some reason for objecting to the term "chemical growth", though the simple fact that rate of increase in dry weight follows an opposite course from rate of increase in wet weight (i.e. "growth") is plainly of importance. There is less reason for objecting to the term "chemical differentiation", for Murray uses it to refer to such entities as percentage of ash and percentage of fat, and shows that, although for by far the greater part of the embryonic period the 552 ON INCREASE IN COMPLEXITY [pt. iii former is descending and the latter ascending, they are doing it in the same phase, i.e. slowly at first and more quickly afterwards. Probably the most significant result which emerges from Murray's symmetrically diphasic plan is that growth-rate and differentiationrate go together, and in opposition to metabolic rate. Enriquez's paper of 1909 contains a foreshadowing of the idea of chemical differentiation (see also Scammon & Ness). S'S. Chemical Processes and Organic Form In considering the processes of differentiation in the embryo, there has been much disinclination to admit their physico-chemical nature. It is the great credit of His that he took the lead in this direction, pointing out that processes such as the production of the neural and amniotic folds were the inevitable results of unequal growth controlled by physico-chemical factors taking place in what was, to start with, an undifferentiated sheet of embryonic cells. His's artificial blastoderms, again, made of pills of dough to which varying amounts of yeast in various places are added, provide a close model for the embryo. D'Arcy Thompson has described the incredulity and opposition which the views of His met with, especially after the publication of his classical letter in the Proceedings of the Royal Society of Edinburgh in 1888. As Garbovski put it, "it is absurd to treat the living being as if it were made up of vesicles, cylinders, and plates, and not of vital units". Embryologists such as Hertwig and Balfour held that, in the study of development, a sufficient causal explanation of one stage had been given when the immediately preceding stage had been adequately described. " My own attempts", said His in a famous passage, "to introduce some elementary mechanical or physiological conceptions into embryology have not been generally agreed to by morphologists. To one it seemed ridiculous to speak of the elasticity of the germinal layers; another thought that by such considerations we put the cart before the horse; and one more recent author states that we have better things to do in embryology than to discuss tensions of germinal layers and similar questions, since all embryological explanations must necessarily be of a phylogenetic nature. This opposition to the application of the fundamental principles of science to embryological questions would hardly be intelligible if it had not a dogmatic background. No other explanation of living forms is allowed than heredity and any which SECT. 3] AND ORGANISATION 553 is founded on another must be rejected — yet to think that hereditywill build organic beings without mechanical means is a piece of unscientific mysticism." Such opposition to mechanical explanations in embryology has of course long been dead, and there are many authors who have put forward such theories to account for the phenomena of differentiation in the embryo. But these theories are for the most part quite physical, depending on physical properties such as elasticity and torsion, so that they cannot be considered in detail in this book. It is also very unfortunate that, owing to the exceedingly small size of the parts undergoing such changes, practically no direct work has been done, and investigators have focused their energies on the preparation of models, made of a variety of materials, such as rubber and plasticine, which can be made, like the yeast pills of His, to exhibit the phenomena of gastrulation and the like. Roux in his valuable review of the technique of " Entwicklungsmechanik " has described such rubber models, and an even greater collection of them is to be found in the paper of Rhumbler. Rhumbler gives in full the literature on this subject (e.g. Morgan; Schaper & Cohen; and Spek), and has many illustrations of invagination models, etc. (see especially his Section F onwards) . These investigations are interesting indeed, but they do not contribute a great deal to our knowledge of what happens in the early stages of embryonic development, although they certainly set forth a number of ways in which it might conceivably happen. Robertson's theory of cell-division, again, in which lecithin was supposed to be broken down at the two nuclear regions of the dividing cell to provide phosphoric acid for new nuclein, and the resulting free choline to diffuse away, reaching its greatest concentration in the equatorial plane at right angles to the line joining the two nuclei, was never shown to hold in actual fact. Like so many of the "Nachahmung" models, it rested upon an unwarranted simplification of the material under discussion. Nevertheless, it was an ingenious suggestion, and, in spite of McClendon's criticisms, it may still be found to contain a modicum of truth, but the reason why it and others like it will not be taken up in detail here is because they are not sufficiently close to the facts. The purpose of this book is to give all the facts that are known about the physico-chemical aspects of embryonic development, and not the theories, which, indeed, would demand a much larger treatise. 554 ON INCREASE IN COMPLEXITY [pt. m That there is at present a certain gulf between physico-chemical research and the "form" and "shape" of the morphologists is a fact which must be faced. In present-day biology, there lives on, still very hale and hearty, the essence of the distinction made by Aristotle between vXt] and eloo<;, matter and form. It is probably at bottom this which inspires the statements so often made by morphologists that do what one will with chemical methods, the meaning of form in animals will always elude one's grasp, for it belongs to another order of existences, a range of concepts intrinsically remote from physics. In so far as the form of living organisms is an expression of a degree of organisation higher than anything with which the sciences of the non-living world have to deal, it is true that we have to deal with something very different from mere heaps of molecules, but crystal form and the colloidal state, which exhibit an intermediate degree of organisation, exist in the inorganic world and can be dealt with by the quantitative methods of physics and chemistry. Examples of the extreme morphological point of view are common; thus Cunningham in 1928, discussing chemical embryology, remarked of two eggs in the same incubator, "Why are the bones formed in one case the bones of a chick and in the other case the bones of a duckling?" For him the fundamental problem is, not how does the rabbit get out of the hat, but why a particular kind of hat should produce a particular kind of rabbit. In one sense the question is simply a special case of the general question, why is the universe what it is and not something else? and thus reduces to a query concerning the fundamentally alogical character of the universe. With such conundrums the scientific investigator is not concerned and Cunningham should have addressed his inquiry to metaphysicians. Cunningham's question had already been raised by C. D. Broad. "The ultimate question", said he, "is, how do these particular material systems called organisms come to have their particular structure or components. So long as we explain their origin by laws, whether mechanical or otherwise, we are merely referred back to earlier collocations of matter, and so on ad infinitum. The explanation in terms of a designing mind on the analogy of humanly constructed machines seems to involve a circle or to end in a mind so different from any that we know that the analogy fails and it is hardly worth while calling it a mind. The explanation by entelechies rests on a confusion and avoids no difficulty which is SECT. 3] AND ORGANISATION 555 raised by the notion of an external designer. The problem, as far as I can see, is extra-scientific and quite insoluble." There is a sense, however, in which Cunningham's question has a meaning for science, and all that can be said, in answer to it is that the exact biologist has a hope, a belief, that in the long run the outward forms and shapes of living animals are as much dependent on the properties of what we call matter, as anything else about them. To object that this is to endow atoms and electrons with occult properties and potentialities is not reasonable, for the only alternative is to abandon all hope of bringing form and shape within the coherent scheme of the scientific world-picture. We cannot resign ourselves to leaving them out in this way, and there is the less cause for despondency when we see what great progress has been made along such lines of research as that pursued by d'Arcy Thompson and summarised in his Growth and Form. Nevertheless there are real difficulties in this subject, and J. H. Woodger, who has acutely felt them, has even gone so far as to maintain that biology may for ever consist of two irreconcilable divisions, morphology and physiology. Experimental and causal embryology, in his view, is only physiology disguised. He would regard morphological description as an end in itself, and the ultimate aim of the morphologist to come down to solid geometry instead of to causal relationships. There may be in this view an element of truth, but in so far as any morphologist holds such an opinion of his goal he must admit himself to be an artist searching for the aesthetic experience of significant form rather than a scientific investigator seeking for understanding of how the thing works — surely the essence of scientific explanation. Another thinker who has vigorously opposed the extension of physico-chemical concepts to include morphology is E. Rignano. In a discussion of the relations of biochemistry with embryology he said, "Now what do the ultra-mechanists do in the presence of these teleological manifestations of the generative and regenerative processes? They direct all their efforts to an attempt to prove that given chemical substances exercise a morphogenetic action on particular developments, hoping to conclude triumphantly that the entire series of morphogenetic phenomena, constituting the ontogenetic development, may be explained completely by physicochemical action. But in this attempt they have mistaken a mere 556 ON INCREASE IN COMPLEXITY [pt. iii release of morphogenetic activities already potentially in the developing embryo for a genuine morphogenetic action". Now it is easy to adduce in the light of biochemical researches, chains of causes which can have the effects seen in the developing embryo. One could start with the experiments of Huxley on the morphogenetic action of thyroxin. We know from Ahlgren's work that thyroxin has definite effects on oxidation-reduction mechanisms in vitro, just as other hormones have, and anything that may locally affect oxidationreduction mechanisms has every chance of affecting the local fatty acid concentration, as is indicated by the work of Hopkins, so that in due course the lipocytic constant or some other such cellular value (cf. Mayer & Schaeffer) will alter and change correspondingly the surfaces of the intracellular phases in that region, with the final outcome that, as in the models of Warburg, one chemical substance may be formed instead of another. These two alternatives may be thought of, for instance, as scleroprotein on the one hand or phosphoprotein on the other, and thence it requires little imagination to picture the most profound morphological changes taking place. These causal chains are being unravelled every day. But Rignano was probably prepared to admit the cogency of these mechanisms within their own sphere; what he wanted to know was, why should one embryo pass through all these changes and come out at the end a dogfish, and another pass through them and come out a skate. The standard answer of exact biology has, of course, been that the genetic constitution of the former animal governs the chemical morphogenetic processes, catalysing this and inhibiting that, so as to produce the results we find. Rignano, Haldane, and their associates have often replied that such a preformationism implies a complexity too great to be imagined when it is faced with the facts of biology as a whole, but this depends on one's imagination. The alternatives, it might be argued, are much worse. The adversaries of genetics too often seem to suppose that every one of the infinite number of characters which they see in a given animal has to be represented in some way, within the nucleus. They forget that large blocks, as it were, of the specific- characters may be the result of single processes set in action by a gene, and do not consider the possibility that a good deal of morphogenesis may be associated with a "delegation of function", the gene activating secondary key-factors, just as statesmen delegate many of their SECT. 3] AND ORGANISATION 557 duties to subsidiary but competent officials. This possibility becomes a probability when the apparently mysterious and arbitrary grouping of characters is considered, colour x always going together with pollenshape jv and so on. In any case there is no need to load all the responsibility for the adult skate and dogfish on to their genetic equipment; any more than we need suppose the constitution of their eggs to differ simply by the presence of two different entelechies, as Driesch would say (see p. 22). It is true that in many respects the chemical constitution of eggs is alike in different animals, but only when very broadly considered. We know that the provision of amino-acids is by no means the same, and serological differences, which have so far only been touched on by a few workers (see Section 19), are likely to be of much importance. Riddle has also emphasised the importance of the environmental factors in contributing to the final result of ontogenesis, so the three principal sets of morphogenetic causes which exact biologists at present accept are thus: (i) The genetic constitution of the egg-cell. (ii) The physico-chemical constitution of the egg's raw materials. (iii) The environment during ontogeny. Woodger observes that modern genetics owes a debt to the preformation theory of the eighteenth century. The old theory identified the "immanent factors" in the egg with the whole of the newly separated individual, and imagined nothing but an increase in volume. The modern theory identifies the immanent factors with certain small bits of the individual, such bits being thought of as related to the qualities of the individual as cause to effect. Woodger considers that we shall get on better by sharply distinguishing between genetics and embryology instead of by attempting to fuse them, as is now the general aim, for he regards the formation of parts as fundamentally or causally separate from the determination of characters. I must refer those who are interested, to the original discussion, but if we are not to treat parts as characters, an entirely new conception of evolution will be required. The difficulty of fusing morphology and physico-chemical biology is, in fact, very real, and Rignano's objections, though they are far from insuperable, are not merely restatements of finite teleology. They bring up the question of how any real epigenesis can take 558 • ON INCREASE IN COMPLEXITY [pt. iii place at all, i.e. how from moment to moment the level of actual organisation in the embryo can rise, a question of much theoretical importance to which I shall return in the Epilegomena. But chemical embryology will never allow itself to be restricted to the description of relatively superficial events in the life of the embryo, such as the appearance of enzymes in the digestive tract. Ii will insist on expanding physics and chemistry, if necessary, to cover the animal level of organisation. It will affirm that if we knew all that there is to be known about the physico-chemical constitution of the egg, we should be able to predict the results of its development. This affirmation does not imply that we should be able to make such predictions before any given case had been observed ; for the fact of emergence is real and it is true that knowing all the properties of simpler systems, including how they could combine with one another, does not tell us how in point of fact they actually do combine. The schemes of science are resultant, not emergent, they cannot describe the complex systems before they have been observed, but they can and do offer reasonable causal explanations of them with reference to their simpler factors, after they have been observed. In this way we are not sure that physicochemical embryology will ever be able to say what a hitherto unknown egg will develop into, but we do expect it some day to be in a position to offer a reasonable causal explanation of the origin of all measurable properties of adult living beings from the measurable properties of their eggs. And form is evidently one of these, just as is physical constitution. There is a great deal of confusion at the present time about such questions and very few workers stop to ask themselves what is the true aim of their studies in causal or exact biology. Moreover, many biologists are uncertain as to the meaning of the facts with which this book is dealing. "I have not so far been able to discern", writes a Belgian embryologist to me, " the truly explicative value of quantitative measures in embryology. Far be it from me to minimise the interest of the analysis of metabolism in development, but I have the impression that it is more thrilling for the biochemist than for the embryologist. For the latter the cardinal problems remain the setting-in-action of development, differentiation, or heredity, and their mystery lies wholly in the laws of cell-life. Now these laws have not so far been elucidated, and we hardly know anything of SECT. 3] AND ORGANISATION 559 how one part of the cell acts on another; thus there remains an enormous qualitative task to accomplish before we can begin to enter on a numerical phase. I have been thinking of all this in reading the monograph of M. Rapkine on the energetics of development, where the physico-chemical facts are presented in a plain sort of manner lending itself well to meditation. But I confess that it all remains sibylline, enigmatic, to me, and some acute friends have admitted the same embarrassment. Where is the explicative value, the light, the link between all these facts so interesting, yet so isolated? I cannot hit on it." This excellent passage, which my friend has kindly allowed me to quote, expresses admirably the doubtful air with which many biologists regard the extension of physico-chemical concepts into morphology. In answer I pointed out that chemical embryology is a very young science and if more attention had been devoted to it in the past, would not show the blank spaces and the gaps of which my friend complained. In a word, Roux's decision to concentrate solely on the revelation of the secondary components" in embryogenesis has had momentous effects, but there were always those who could not resist the temptation of going deeper down to the "primary components" before all the mass of facts had been dealt with on the more superficial level. Among such miners at the deeper levels are the physico-chemical embryologists. It is true that an enormous qualitative task remains to be accomplished, but there is room enough for both kinds of work and it will surely be through the close co-operation of both kinds of worker that the facts will lose their sibylline and enigmatic character. 3-4. The Types of Morphogenetic Action In order to have some concrete idea as to how the processes of morphogenesis may be supposed to go on, and to form a background for chemical facts, one may construct a table in which are placed all the different kinds of process that may go on during differentiation. Such classifications have been made by KeUicott and by Jenkinson, and tables constructed from them are given here. (Charts IV and V.) Jenkinson's table may be taken first. The movements of single cells which he enumerates are of much importance. The lower-layer cells of the blastoderm of elasmobranch fishes, for instance, are free and > < <.2 23 o o <u be; ■rt g

-■ G ^ < s"? -C-«  bD H" _g +-» a i/5 73 a «  u c ^ ■^■5 -^g^ < o ►J w S Q. > H < 0) 2 O -0. a. o w ■^8 O 1^ he. 23 Ph O a cS H on cell-division, as in the work of Reding & Slosse and of Vies, Dragoiu & Rose may, liowever, be mentioned in this connection. The two tables should be borne in mind throughout all succeeding chemical discussions. Just as perforations may be probably due to local controlled autolyses, so general disappearances of masses of cells may be due to the same factor operating on a slightly larger scale. This conception is made all the more Ukely by the work which has been done since 191 9 on insect metamorphosis, and which goes to show that there also an autolysis plays a large part. Again, the changes in the form and size of individual cells which may grow larger or smaller may probably depend on the increasing or decreasing concentration of storage materials in them, and here there is an extraordinarily wide field for histochemical methods. Unfortunately these are at present so unsatisfactory that it may be a long time before this aspect of morphogenesis can be even approached. The separation of cells and their fusion into surfaces and layers may also be due to local variations in the sterol and lipoid fractions (tissue constants), SECT. 3] AND ORGANISATION 565 the molecules of which might, by changing either their constitution or their position in space or both, be the underlying factors responsible for the process. An immense field lies open here for the application of micro-estimation methods on sections of tissue dissected out from embryos in the early stages of development, a field which has not so far been entered by one investigator. Processes such as chemotropism, moreover, probably have much importance in morphogenesis. This has been emphasised by Kappers whose work on neurobiotaxis has thrown much light on the early development of the nervous system. Associated with this conception is that of the localised presence of enzymes. Everything, indeed, depends here on the word "localised". The localisation of substances in the embryo is a study which has not yet begun, for nearly all we know about its physico-chemicallife is concerned with it as a whole and with its immediate surroundings. The study of embryonic hormones, for instance, to which much attention will later be given, has not so far been carried on from this point of view, although it does seem clear that the embryo in the early stages is quite devoid of "chemical messengers", and only acquires them later at definite time-points in its development. One may say that the "Nachahmung" school have done much to suggest possible morphogenetic processes without doing anything to find out what they actually are in point of fact. The only means by which this can be done are applications of exact physico-chemical methods, and these will have to be made use of more and more in the future. We may, however, before leaving this subject take one instance where chemical work has thrown a light on the state of affairs in the young embryo. In 1926 I studied the free and combined carbohydrate contained in the embryonic body of the chick embryo from the 5th day until hatching. Thus 100 gm. of embryo (water and soHd together) contained on the 5th day of development 160 mgm. of total carbohydrate and 8 mgm. of free carbohydrate. During development the former falls to about 100, and subsequently rises to over 300; the latter rises all the time in an S-shaped curve to about 48 mgm. This would indicate that the carbohydrate present in the embryo on the 4th and 5th days, at which time there is more in proportion than at any subsequent period, is not in the form of free glucose, and it is certainly not glycogen. The dry weight data gave even more striking results. 100 gm. of dry embryo contain on the 5th 566 ON INCREASE IN COMPLEXITY [pt. iii day 3000 mgm. of total carbohydrate and about 100 mgm. of free carbohydrate. After that point the total carbohydrate continuously falls, reaching a value of 1750 on the i6th day, while the free carbohydrate continuously rises, its highest point being reached on the nth day with a value of 360 mgm., after which it falls, but not below 220 mgm. We have therefore a considerable proportion of the sugar present throughout not free and not in the form of glycogen. It is very probable that this fraction can be called "mucoprotein glucose", and that it is present in the form of combination with a protein molecule. Its quantitative relationships are interesting, for it amounts to 2550 mgm. per cent, dry weight on the 5th day, but by the nth day it has fallen to about 1000 mgm. per cent., after which it remains about steady at that level. No stress is laid on the absolute magnitude of the figures, for the best methods we have for estimating total carbohydrate almost certainly give high results, but there is no reason for supposing that the relative values are not significant. It seems that this behaviour can be correlated with facts that have long been known to histologists. Von Szily was the first to describe a cell-free connective-tissue or fibrous ground-substance filling up all the cavities of the embryonic body in the early stages. This has some affinities with the cardiac jelly of Davis, and Baitsell has recently examined its properties with the aid of a micromanipulator. An account of its comparative histology is given by Biedermann. It appears to be secreted by the cells, and provides them with a homogeneous matrix, a kind of natural culture medium in which even migration may take place if and when it may be necessary. As development proceeds, the substance does not disappear, but becomes relatively less and less important in relation to the body as a whole. The nearest equivalent to this ground-substance elsewhere is the Wharton's jelly of the umbilical cord and the vitreous humour of the eye. It is significant that both these tissues are known to be very rich in substances of the mucoprotein type, and the importance of mucoprotein in the beginning of embryonic life leads inevitably to a correlation with the jelly of von Szily. It explains, or is explained by, the high proportion of combined glucose other than glycogen at that period. This example is intended to give an idea of the kind of correlation between chemical and morphological characteristics which is necessary for the embryology of the future. SECT. 3] AND ORGANISATION 567 Such correlations, in the position in which we now are, are difficult. It is also very difficult to bring together the work of the "Entwicklungsmechanik " school and the results of physico-chemical work, but close attention must be given to it in order to orientate the point of view adopted in this book. Any exhaustive discussion of it is rendered unnecessary in view of the papers by Huxley and Spemann, and the excellent book of de Beer. Other recent reviews of experimental embryology are those of de Beer; Przibram; Mangold; Brachet; Weiss; Gilchrist; Hogben and Schleip. It will not be possible here to do more than indicate a few of the outstanding problems to which modern experimental embryology has supplied tentative but fairly certain answers. 3-5. Pluripotence and Totipotence One may first ask, therefore, what answer has been given to the question of organisation in the unfertilised egg-cell, or, in other words, what modicum of preformationism are we still obliged to admit into our conception of the undeveloped egg. In the very earliest stage of development two important critical points occur, or two stimuli are received, the first causing the egg to become radially symmetrical about an axis, and the second causing it to become bilaterally symmetrical in any one of an infinite number of planes passing through that axis. What these processes actually are is entirely unknown, although we can make a guess, as will later be seen. But the fact that they have to occur before differentiation and growth can begin makes it impossible to speak of the spatial arrangement of already differentiated material in the egg, which was the basis of the classical preformationism. It is now correct to say that heredity does not account for the individual but for a number of potentialities, some of which are brought into being in the individual. The only predetermination which exists is the assurance that, if the potentialities in question are brought into actuality, they will produce an organism belonging to the same species as its parents. The appHcation of chemical methods to the undeveloped egg will probably thus in the future not be so histochemical as would have been the case if an arrangement of material had existed there. Nevertheless, what has been said apphes to the vertebrate egg only; there are invertebrate eggs, such as those of Cerebratulus , which do seem to have an intracellular determinate arrangement. The state 568 ON INCREASE IN COMPLEXITY [pt. iii ment here made may be differently expressed by saying that preformationism is only true in the sense that the constitution of the fused nucleus must have reference to, though it dots not actually resemble, the adult form. Protoplasm, and yolk, on the other hand, as Lillie on Chaetopterus, Morgan & Spooner on Arbacia, and Morgan on Cumingia have shown, are homogeneous as far as this goes, since normal development can take place after centrifugation and displacement of visible constituents. (See p. 346.) It is known that eggs vary a great deal as to their dependence on the arrangement of the initial material in the egg-cell. The extent to which they do so is more or less ascertainable by the extent to which they exhibit totipotence and the rapidity with which they lose it. The phenomenon of totipotence was originally discovered by Driesch in 1900, who found that, if the blastomeres of the seaurchin's egg were separated at the 2-cell stage, each nevertheless gave rise to a perfect larva of half the normal size. If the blastomere was taken from the 4-cell stage it would produce a larva of a quarter the normal size. Blastomeres from the 8-, 16- and 32-cell stages did not gastrulate well, and not at all if they were derived from the animal pole. No blastomere from the 64-cell stage will produce a gastrula. However, although one blastomere from the i6-cell stage animal pole will not gastrulate, four together will, and will go on to produce a perfect larva, so that the inability cannot be due to lack of specific enteron-forming substance, but simply to too small size. The gradual loss of totipotence, therefore, is an index of the degree to which the egg depends on the localisation of organ-forming materials in the original egg-cell. In some animals the blastomeres are able to regulate their development, if isolated, and produce a perfect larva, in other animals they are not: the former are termed "regulation eggs" and the latter "mosaic eggs". Examples of the former group, besides the echinoderm eggs just mentioned, are the coelenterates, which, according to Zoja, can be totipotent up to and including the 4-cell stage, and the nemertines; but, in this latter case, blastomeres from the 4-cell stage are not perfectly totipotent, but give rise to larvae lacking a few of the usual parts (Zeleny) . The group of regulation eggs passes over without break into the group of mosaic eggs. Thus in Amphioxus, totipotence exists only at the 2-cell stage, as was shown by Wilson's experiments, and the same applies to amphibia, but only under certain conditions, i.e. if each blastomere SECT. 3] AND ORGANISATION 569 contains some of the grey crescent region, which in time will develop into the dorsal lip of the blastopore (Brachet) . With the ctenophora we pass to a purely mosaic egg, for, as Fischel has shown, though the normal ctenophore has eight combs, the result of a 2-cell stage blastomere is one with four combs, that of a 4-cell stage blastomere is one with two combs and that of an 8-cell stage blastomere is one with only one. Obviously, then, something was distributed more or less evenly among the blastomeres as the cleavages took place, and cleavage of the cells involved equal cleavage of the comb-forming substance. Crampton has discovered very similar phenomena in molluscan eggs {Ilyanassa), Wilson in those of Dentalium, Driesch in those oi Myzostoma, Stevens in nematodes, and Conklin in ascidians. Loss of totipotence may even take place gradually before fertilisation. The organ-forming materials may be actually visible in the undeveloped egg, as, for example, the brownish yellow pigment in Styela eggs (an ascidian) which was found by Conklin to be necessary for the formation of muscle. On the other hand, in some cases, as Morgan has shown, pigments and formed elements are present which, while normally distributed to certain cells, can be centrifuged into a different position in the egg without in the least affecting development. Other examples of this have already been given. In 1926 Duesberg found that some of the elements which seemed to be necessary for development, and could not be centrifugally displaced without damage, were mitochondria. But the most significant experiments were those of Jenkinson, who found that normal morphological differentiation could go on in the frog embryo even after the egg had been centrifuged and the fat globules had been made to get into the "wrong" cells. In this case the brain was formed quite normally, though it contained, at any rate as far as histological examination could show, a much greater quantity of fat than usual. Although this process could not be carried beyond a certain point, it did show that fat alone was a structural material rather than a specific organ-forming substance. These regulatory processes occurring after centrifugation were shown by Konopacka for the frog's egg to be remarkably independent of temperature. Totipotence phenomena probably underlie the remarkable cases of Polyembryony which occur sporadically throughout the animal kingdom (cyclostomatous Bryozoa, earthworms, parasitic Hymenoptera, and Armadillos) and which have been reviewed recently by Patterson. The 570 ON INCREASE IN COMPLEXITY [pt. iii Armadillo produces four identical quadruplets from one egg, and a single Ggg of the parasite Litomastix truncatellus gives rise to about 1500 individuals. 3*6. Self-differentiation and Organiser Phenomena The passage of time between fertilisation and hatching can, it has been found, be divided into three periods. These may be summarised thus: (i) division and re-arrangement of pre-existing material and structure (up to the end of the formation of the germinal layers), (2) primary or non-functional differentiation in which the organ rudiments are determined irrevocably and their early differentiation proceeds, and (3) secondary or functional differentiation, in which the inception of active function by the new organs brings about important consequences. "Up to 1910", it has been said, "the principal achievements of experimental embryology could be summarised thus: first, that nuclear division in early development was not differential, the nuclei of the embryo all being equivalent; second, that most fragments of the germ could, before the onset of gastrulation, if of sufficient size, regulate themselves to produce a whole embryo or an approximation to it; third, that in a few cases definite 'organ-forming materials' existed in the fertilised &gg — visibly differentiated regions causally correlated with the development of certain organs; and fourth, that Roux's doctrine of the struggle of the parts was valid in the later stages." The intermediate part of development was not, in fact, very well understood. But Spemann's discovery in 1918 that the region of the dorsal lip of the blastopore was a differentiator or organisator, as it was called, threw a new light on the middle period. By grafting a piece of it into another embryo (of different species so as to recognise the parts by their colours) he found that it would cause the host tissues between it and the animal pole to form the primary axial organs, i.e. neural plate, notochord and somites. It would initiate, in fact, the selfdifferentiation process, and would cause cells which previously would have had the potentiality of becoming almost anything according to their position to set out on their path of irreversible differentiation. Then Brachet showed that the organiser in the dorsal lip region exerted its formative effect behind as well as before. If a cut was made so that continuity between the organiser fragment and the posterior region was interrupted, there would be no formation of SECT. 3] AND ORGANISATION 571 notochord and somites behind this obstacle, and, on the other hand, a cut between the organiser and the anterior region would exert the same effect there. The influence evidently demands continuity for its operation. It is important to distinguish the actual differentiation of this middle period from the invisible determination which foreruns and controls it. Spemann showed, for instance, that in the early neural plate stage, before any sign was present of the optic vesicle, there was a predetermination of the eye as a whole, and also the separate areas for optic stalk, retina, and tapetal layer. Bulbar, ventricular, atrial, and sinus substances are similarly determined before any trace of the heart appears. Huxley proposed the term "chemodifferentiation " for this invisible point of decision as to the fates of cells *. After this point the organism is a chemical mosaic of qualitatively unUke regions, in which regeneration is impossible. "If a gastrula be cut into two", he says, "each half forms only those organs which it would have formed as part of the whole; again, if a piece of the future brain region is cut out of the embryo in the neural plate stage and then grafted back in the reversed position, its different parts still produce the structures that would have been produced in its normal position. Each chemically determined region is separate and irreplaceable." And this state of affairs proceeds until the onset of function, after which, with nervous control and the activities of hormones in humoral channels, a third and quite different period comes into being. It is most important to note, meanwhile, that the point of chemodifferentiation is associated with gastrulation. In the section on resistance and susceptibility a good deal of evidence will be presented which tends to show that gastrulation is one of the main critical points in development, points at which a less degree of external interference is required than at any other, in order to make development go wrong. The opportunities afforded by the tissue culture technique have also been utilised in the study of the middle period of self-differentiation or irreversible morphogenesis. Obviously the cultivation of an explant of a certain region of embryonic tissue at different stages will (or may) reveal whether it has been determined or not. This deter

We do not know how far this process is really chemical. Decision would be easier if we knew the rate at which, say, medullary plate ectoderm becomes specifically eyeforming substance. Determination-rate is probably at least as important as Growth-rate and Differentiation-rate. 572 ON INCREASE IN COMPLEXITY [pt. iii mination may from one point of view be regarded as the point where the curve of decreasing totipotence reaches the base-line. For totipotence persisting into later stages, such as blastula and gastrula, demonstrates its presence by the capacity of any piece of tissue to develop into structures which it never would have done if it had been left to itself The name usually given to this first-period indeterminateness is "plasticity". Thus before gastrulation Spemann found that a piece of epidermis normally destined to fold in to form the ectodermal nerve-cord can be exchanged for a piece of ordinary epidermis with the result that the presumptive nerve-cord will turn into skin and the presumptive skin into nerve-cord. Such plasticity occurs all over the embryo before gastrulation, and Spemann was able, in some very beautiful experiments, to work with the darkcoloured Triton taeniatus and the light-coloured Triton cristatus together so that dichromic grafted embryos were produced. But if one attempts to make such transformations as these after gastrulation has been completed, one meets with no success whatever — even though no differentiation at all is visible, the hidden process has determined the fate of the cells. In the preliminary plastic cells even the germ layers are interchangeable. Pieces of presumptive ectoderm can be planted just beneath the dorsal lip of the blastopore, and being carried in as invagination proceeds can afterwards be recognised as taking part in the constitution of mesoderm or endoderm. In other words they behave "Ortsgemass", i.e. according to the arrangements of their host, and not "Herkunftsgemass", i.e. according to the arrangements of their origin. Such pieces can be earmarked by the use of vital stains, as Goodale in his work against the concrescence theory was the first to show, and Vogt has in more recent years done many experiments of this nature. The converse, namely, that presumptive endoderm or mesoderm can give rise to ectoderm, has been demonstrated by Mangold, and similar remarks hold with respect to the limb buds. The fact that the organiser of one kind of Triton can bring about its effects in another kind of Triton naturally led to the question of what taxonomic distance was necessary to prevent the action of the organiser. Geinitz, working on this point, found that the organiser of even very different species such as Amblystoma or Bombinator could be grafted into Triton embryos with efficient results. The grafts might SECT. 3] AND ORGANISATION 573 even come from different genera, families, or sub-classes. Obviously the region of the dorsal lip has in different embryos something very much in common, which can act in a most remarkable way on neighbouring tissues. More extraordinary still, the influence is not a peculiar property of the blastopore lip, for other tissue can be grafted into it, and then can be used as an organiser after having been, as it were, "infected". We have already seen that the organiser requires contact for the spread of its influence and will not function across a cut. The spread, Spemann found, also takes a measurable time, for at an early stage the ectoderm near the organiser is determined to turn into notochord, while further away it is still indifferent. Ruud & Spemann threw light on the finding of Wilson already referred to when they observed that, if the newt gastrula is divided into halves, only that half which contains the organiser will develop. Brandt has thrown further light on the process of chemodifferentiation and has divided it up into three periods or phases, reversible, critical, and irreversible. Thus different amphibian species which look externally in the same developmental stage may or may not be equally chemodifferentiated. The influence that radiates from the dorsal lip is not apparently under the control of any other internal factor; it is therefore called an "organiser of the first grade". But some of the organs which differentiate under its influence show in their turn an organising action on the structures developing near them; thus, the axial structures exert an effect on the development of the thoracic and abdominal viscera [situs inversus viscerum experiments). They are therefore called, as Spemann suggested, "organisers of the second grade". In the same way the eye-cup is usually, though not always, an organiser of the second grade with respect to the lens, as Spemann; Lewis, and many others have shown. There is certainly no doubt that some organs at a certain stage contain within themselves all the factors necessary for their complete development as far as the beginning of the third period, while others do not, and seem to depend upon adjacent organs. So far only a few of these relationships have been worked out, but it is very probable that the factor of time is an important one, and, while at one stage an organ or part may be dependent upon some neighbouring second-grade organiser, at another stage it may be quite self-differentiating. Examples of these relationships can be found in 574 ON INCREASE IN COMPLEXITY [pt. iii the work of Streeter on the amphibian ear, which showed that the membranous labyrinth is self-differentiating, and of Luther, which showed that, on the contrary, the cartilaginous auditory capsule is dependent-differentiating, i.e. in relation to the auditory vesicle as to a second-grade organiser. Again, in Bombinator the visceral rudiments are self-differentiatory, according to Holtfreter. Tissue culture methods have been already referred to as important in the analysis of the differentiatory power of a given anlage. These may be of various kinds. One of the most valuable is the transplantation of the tissues of bird embryos on to the chorio-allantoic membrane of another c^gg, where it can easily be seen whether or not they continue to develop in isolation, and, if they do, what they develop into. In this way Danchakov found that the blastoderm of the chick develops properly in isolation, but only after some hours. The degree of differentiation here depends entirely on the age of the embryo from which the piece is taken. Thus Hoadley found that pieces of an embryo only 4 hours old will, when transplanted on to the chorioallantoic membrane, produce an eye with pigment cells only; if the age of the embryo is 6 hours, an eye with pigment and retinal cells will be produced, and, if the age of the embryo is 8 hours, an eye with pigment and stratified retina will result. At 10 hours the primitive groove is formed, and, if a piece be taken from an embryo after the beginning of somite formation, a completely self-differentiating eye will develop on the chorio-allantoic membrane. Other workers have made similar experiments with different regions of the chick embryo as follows: limbs and limb girdles — Murray & Huxley; Selby & Murray; and Spurling; eye — Barfurth & Dragendorff; head — Lillie, and Danchakov; spinal cord — Agassiz & Danchakov; nose, ear and mesencephalon — Hoadley; metanephros — Atterbury; otocyst — Fell; gonads — Minoura; mesonephros and primordial germ cells — Humphrey; various tissues — Hiraiwa; heart — Danchakov & Gagarin. The other main form of explantation work is the in vitro technique of tissue culture. This has been used to study self-differentiation by Reinhov, who found that the kidney of an embryonic chick would develop glomeruli, tubules, and capillaries in vitro, and by Strangeways, who got good differentiation in embryonic cartilage. Working with the same technique Strangeways & Fell showed that the limb buds of embryonic chicks have only a small power of self-differentiation in vitro (or in vivo in trans SECT. 3] AND ORGANISATION 575 plants), but that, on the other hand, the eye showed remarkable capacities, producing all the constituents of the normal eye, but not growing much in size. And with 6-day femurs, perfect self-differentiation was produced in vitro by Fell & Robison. Most interesting of all, Danchakov found, while studying the degree of self-differentiation in various chick tissues, that the mesonephros degenerated after a time in the grafts, just as it does in the intact embryo. A commentary on the relations between tissues acting as second-grade organisers and the tissues they influence is seen in the observations of Champy, who found that epithelium and connective tissue (from post-embryonic stages) in pure cultures by themselves underwent dedifferentiation,iand reverted to an embryonic condition. But if they were both present together in one culture the differentiation was maintained, yet only so long as the connective tissue constituent of such a culture was living. If it died the epithelium dedifferentiated. In just the same way the kidney tissue of the mouse grows alone as a sheet of undifferentiated Itissue, but can be made to take on its usual character by the addition of connectivetissue cells to the cultures (Drew). "It will be seen", says Huxley, "that the discovery of the organiser and of the gradient coordinate system enables us for the first time to give a coherent formal account (however imperfect in detail) of the early stages in development. In so doing we sail clear of the difficulty which has beset so many minds of understanding how differentiation can be compatible with absence of qualitative nuclear division. Loeb in his book The Organism as a Whole was driven to assume that the Mendelian chromosomal genes were only responsible for minor characters, the main course of development being determined by the ovum, which, owing to its assumed possession of organforming materials, was to be regarded as the 'embryo-in-the-rough'. It is now seen that the egg cannot be held to become the ' embryoin-the-rough ' until chemodifTerentiation has started. After this moment specific organ-forming substances are all-important, but in most unfertiHsed eggs they scarcely exist. The production of these organ-forming substances depends upon the varying interaction of organiser and genes in regions of various activity. The differential which determines the variation of activity is the system of metabolic gradients, which, although definitely organised, is very far from constituting the egg an embryo, however much in the rough. One 576 ON INCREASE IN COMPLEXITY [pt. iii of the two main gradients is determined at or after fertilisation by agencies external to the ovum. The other is determined in the unfertilised egg, but from an analogy with other forms, it is to be expected that this too will be found to have been determined earlier by agencies external to the oocyte (position of the oocyte in the germinal epithelium, blood-supply, etc.)." Such is the general scheme to which the recent researches in experimental embryology have led. It is to be remarked that it includes in intimate association with the working of the organiser the theory of metabolic gradients. It will be necessary to treat this in some detail since it is based on biochemical conceptions, and in some cases biochemical experiments. But, before doing so, some further space must be devoted to the organiser. What determines the direction of the differentiation induced by an implanted organiser? It might determine for itself the direction in which it would radiate its influence or this might be a function of the host. Geinitz's experiments have shown that the latter alternative is the more probable one. At exactly what stage the cells of the dorsal lip acquire their organising power is naturally very difficult to ascertain, and has not yet been ascertained with accuracy. Moskovski long ago found that, if the grey crescent in the frog's egg was injured by pricking with a hot needle or other means, no embryonic anlage or organ-rudiments would be developed, although gastrulation was not inhibited, and Spemann obtained precisely similar results by eliminating the dorsal (i.e. grey crescent-containing) blastomere at the 2-cell stage. The organiser must therefore originate very early. But Spemann has so far not been able to get a transplanted piece of grey crescent to function as an organiser, although Bautzmann reports success with parts of blastulae. Of second-grade organisers Spemann says, "The optic vesicle can be regarded as the organiser of the lens. Yet the vesicle itself, as well as any organising power which it may possess, is not evolved through mere self-differentiation of an anlage of an early gastrula: on the contrary, the differentiation of this particular anlage itself is determined by an external stimulus. We have seen that at an early stage of gastrulation the presumptive eye-anlage can be replaced by presumptive epidermis and it would be possible to choose the implant so that it should contain the presumptive lens area. In this case the two ectodermal layers would exchange their roles — the stimulator SECT. 3] AND ORGANISATION 577 becoming the stimulated. The optic vesicle may then be termed from this point of view an organiser of the second grade". The same state of affairs exists between the medullary plate and the roof of the archenteron. Spemann showed that a piece of presumptive ectoderm, e.g. a piece of presumptive medullary plate, could be made to develop into archenteron when implanted into the dorsal lip of the blastopore and allowed to pass inwards with the gastrular invagination. But, on the other hand, the same experiment can be reversed, and a piece from the roof of the archenteron taken, grafted somewhere else and so allowed to induce a medullary plate in any indifferent epidermis. "Geinitz", said Spemann, "recently combined these two experiments into one. A piece of presumptive epidermis was removed at the beginning of gastrulation from an embryo of Triton taeniatus, heavily stained with intra-vital stains and implanted into the dorsal lip of the blastopore of another unstained embryo at the same stage of development, in such a manner that it invaginated and formed a portion of the archenteron. It was then removed again and this time transplanted into the cleavage cavity of a third embryo at the onset of gastrulation. In the course of gastrulation it came to lie under the ectoderm and induced in the latter the formation of a secondary medullary plate. This transplant, if left in its normal environment, would have differentiated into epidermis ; under the influence of the archenteron — ^into medullary plate, perhaps into eye-anlagen and would then have induced the formation of a lens. In its third environment it became itself an organiser and induced the formation of a secondary medullary plate." What is the extent of the region occupied by the organiser? This question has been investigated by Bautzmann, who definitely settled it by transplanting small trial pieces from the whole surroundings into the blastocoele cavity of other gastrulae. Gastrulation brings these under the ectoderm and then their power of induction, if they have any, will manifest itself, as Marx showed, by the appearance of a medullary plate in the overlying ectoderm. The region of the gastrula, then, which may be said to contain the organiser, is a semicircular area above and beside the upper lip of the blastopore. Normally it is invaginated, and gives rise to notochord and mesoderm. A good deal is known about the structure of the organiser region. It must have an axial longitudinal structure which is not lost by transplantation, for the second embryo may readily be obtained NEI 37 578 ON INCREASE IN COMPLEXITY [pt. iii at right angles to the host embryo if the organiser is transplanted into it at right angles. It must also have some sort of "laterality", as has been proved by Goerttler *. Roux long ago concluded from work with half-embryos that each lateral half could develop more or less independently of the other, and Vogt, by subjecting salamander eggs to a temperature gradient by holding them in a silver plate on each side of which water at different temperatures was circulating, found just the same thing. The left half could be made to hypertrophy in relation to the right half or vice versa. Huxley made very similar experiments with a less abrupt temperature gradient. These experiments showed that the embryo as a whole has a half-structure. But Goerttler went further in removing the left side of a dorsal lip and grafting a right side, as it were, from another embryo into it; whereupon two right halves of a medullary plate and two right medullary folds were formed. Thus the organiser itself must possess laterality. It has also what might be called a regional structure, in that different parts of one organiser tend to produce different embryonic structures. But the question is an extremely complicated one, for a given part of the organising region tends to induce more than it ought, and there are signs that the term "harmonious equipotential system", which was first applied by Driesch to totipotent blastomeres, may also be required for certain parts of the organiser region at certain times. Of the physico-chemical nature of the organiser very little is known. The small size of the amphibian material on which most of these experiments have been made, and the body of experimental difficulties as a whole, have made this side of the subject very backward. Heteroplastic transplantations have shown, as we have seen, that the organiser is not species-specific. It cannot act across a gap, but requires continuity of the cell-mass for its effect. The inductive power of the cells of the dorsal lip is not abolished by drying them, according to Spemann, but freezing and thawing does lead to its loss. Mangold has begun some interesting experiments on the quantitative aspects of the working of the organiser. Whatever is the mechanism of the organiser, there is evidence that it retains its activity unimpaired for a long time. Thus Mangold

Yet Spemann has found the inductive power still present in squashed pieces from the organiser region. Does this mean that the laterality of the organiser is stereochemical rather than cytological? SECT. 3] AND ORGANISATION 579 found that a piece of the brain of a free-swimming larva would still induce a medullary plate in the early embryo, so that the organising power was still present, although it had long been unnecessary. Again, Bautzmann transplanted a piece of notochord from a neurula into the blastocoele cavity of a much younger embryo, and found that the notochord fragment induced a medullary plate in the ectoderm above it. It must, then, have retained the power to do so long after it had been necessary to exercise it in the process of normal development. Precisely analogous are the experiments of Wachs, who found that regeneration of the adult amphibian lens takes place under the inducing influence of the retina, just as had happened originally in ontogenesis. Such experiments demonstrate that the organiser persists into post-embryonic life. Spemann & Mangold also found that, when they transplanted a piece of medullary plate into the blastocoele cavity of an ungastrulated embryo, it would induce in its turn another medullary plate. This process, which they called " homoiogenetic induction", is really a special case of the action of a second-grade organiser, in which a tissue produces a replica of itself. A process occurring in the first and second periods of embryonic development, which has not so far been touched on, is that of "double assurance". Thus the eyeball in amphibia may induce a lens in foreign epidermis, but in some amphibia the lens may develop on its own in the absence of any eyeball, i.e. is self-differentiating, and not dependent on the action of the optic vesicle in its capacity of second-grade organiser. In Rana esculenta both faculties have been conclusively shown by Spemann and Filatov to coexist. Probably further analysis of development will show that this double assurance principle plays a great part in morphogenesis, and that cells only become what they do under the influence of as many as three or four contributing causes *. The double assurance principle may correspond to the factors of safety which appear in structural engineering, so that, if one process goes wrong, the embryo can still manage to complete its development with the aid of the others.

Thus Bautzmann got differentiation of medullary plate from fragments of Triton blastulae which lacked the entire organising region. Similarly Hoadley got selfdifferentiation of parts of the chick blastoderm at 4-6 hours, whereas Waddington showed such parts to be still plastic up to 18 hours or longer. It is clear that chemical determination is often controlled by more than one agency, and that the times of activity of these agencies may overlap. 37-2 58o ON INCREASE IN COMPLEXITY [pt. iii It is to be noted that all the experiments which have been described have been carried out on amphibian material, but the evidence which leads us to see in them a validity over all types of embryo, even mammalian ones, is rapidly accumulating. Seidel has extended the concept to insect eggs, while von Ubisch ; Runnstrom ; and Horstadius find evidences of it in those of echinoderms, Wilson in those of annelids, and Graper; Hunt and Waddington in those of birds. It is highly probable that organiser phenomena will in time be found to exist in all varieties of embryo. The division of embryonic development into three main periods, however, is more generally certain, and may be taken to hold in all cases. As we have already seen, the junction between the first and second periods of development varies with different eggs. In some (regulation eggs), the point of chemodifferentiation does not occur till gastrulation — this is its latest point — but in others it occurs earlier, at some time during cleavage and blastula formation, while in pure mosaic eggs, of which few are known, it occurs before fertilisation. The duration of the second period, the period during which irrevocable differentiation is going on, is rather variable. It ends with the beginnings of function on the part of the foetal organs. 3-7. Functional Differentiation In this third period further differentiation may be dependent on functional activity for its proper progression. The classical example of these mechanisms is the circulatory system, in which Oppel & Roux found that the structure and constitution of bloodvessels depended largely on how they were being utilised by the circulation as a whole. Fischer & Schmieden, for instance, transplanted a section of vein into the course of an artery, where it acted perfectly well, but took on the characteristics of an artery, i.e. its connective tissue content increased and its muscular walls were more than doubled in thickness. Exactly the same thing happens with regard to the central nervous system, and to the bones. "The formation of the normal structure of the bones ", concluded Landauer, "is caused largely by the static conditions of muscle tonus during embryogeny." Diirken's well-known experiments may be mentioned, in which, when the hind limb buds of a frog embryo are removed, the hind brain does not develop normally. Again, Babak observed that in tadpoles the area of the active intestinal absorptive SECT. 3] AND ORGANISATION 581 surface is directly proportional to the amount of vegetable matter in the diet; if it is large, the intestines are capacious and long, if it is small, the intestines are small. In this connection the case of the changes in the intestinal tract of the opossum studied by Heuser is of interest, for it has a gestation time of only 1 3 days, and has to live some time on milk in the pouch. Babak's work was criticised by Klatt, but has been confirmed more fully by Elven. The three periods in the life of the embryo are now known to have different relations to regeneration. Przibram's "law of apogenesis" holds true in the main, namely, that the younger an animal the greater are its powers of regeneration (cf. the work of Abeloos on Planaria dorotocephala) . But this must be qualified by the statement that, during the intermediate period, no regeneration is possible. Each individual part and organ of the embryo has its work cut out to differentiate into its destined form, and any replacement of lost parts cannot be made. Thus Spurling found that the limb buds of the chick cannot be reformed during this middle period. But in the later period of functional differentiation, regeneration is possible, as has been shown by Olmsted; Morgan & Davis; Davidov; and Nussbaum & Oxner. Perhaps the best study of this is the work of Mackay, Mackay & Addis on compensatory hypertrophy after unilateral nephrectomy. Their figures were as follows : Age in months % hypertrophy I 52-6 3 36-7 6 32-8 12 32-2 It is in this later period of functional differentiation that Roux's doctrine of the struggle of the parts has its significance. The signs of this equilibrium between organs and tissues do not emerge except when the organism is subjected to the stress of an unfavourable environment. Then some parts will be found to have the preference and to take a relatively greater share than the others in the available food-supply. The reproductive glands have an important position here, e.g. the testis of the starved rat in Siperstein's work, and there are many instances where the ova in course of preparation draw to a great extent upon the remainder of the body. A discussion of these facts will be found in the appendix on the maturation of eggs ; here it may suffice to mention the work of Greene on the salmon, who found a remarkable constancy in the chemical composition of the ovaries, 582 ON INCREASE IN COMPLEXITY [pt. iii although that of the rest of the body was varying considerably according to the food eaten. 3 -8. Axial Gradients It is now time to turn to another aspect of the analysis of embryonic development, the theory of axial gradients. A convenient transition is afforded by Bellamy's discovery that the dorsal lip of the blastopore and the animal pole in the frog's egg are regions of "high protoplasmic activity". According to the theory of axial gradients, this would mean that the metabolic rate at those points — obviously of great importance as being the seat of the organiser — was higher than anywhere else in the embryo at that moment. It is plain that this is a matter of much interest, and it is therefore necessary to examine in some detail the theory of metabolic, axial or physiological gradients as a whole. Much of the evidence on which it is based is to be found in the four books of Child, and Abeloos and Ranzi have written valuable reviews of the subject. The fundamental conception lying at the base of these views is that of axiate pattern. Child emphasised in all his work the idea of polarity and of gradients of activity between poles. He proposed that we should think of the embryo or the animal as existing in a three-dimensional graph or co-ordinate system, and being constituted in a kind of pattern of axes of symmetry. Each axis would pass from one pole to another, but along its length the protoplasmic activity would not be constant; on the contrary, it would be very high at one pole and very low at the other, dwindling away at a definite and measurable gradient. Such an axis may or may not at its origin in time have a visible morphological outward sign of its existence. It may or may not last as long as the differential growth to which it gives rise. In fact, ontogenesis from this point of view is the clothing of the original protoplasmic axiate pattern with a corresponding morphological axiate pattern. The anatomical gross differentiation of parts with reference to a given axis is preceded, as it were, by the appearance of a gradient of physiological activity along this axis. By a gradient of physiological activity Child meant a series of quantitative differences between the properties of the cells, following a definite orientation with reference to the eventual pattern of the animal. Child may be said to have transferred those co-ordinate diagrams which d'Arcy Thompson used to SECT. 3] AND ORGANISATION 583 demonstrate animal form from appliances of convenience into descriptions of actual fact, and to have substituted for morphological axes, axes of physico-chemical difference. Difficulty was bound to arise when some sort of identification of the physico-chemical differences was attempted, and Child, content with a loose and general association, chose metabolic rate as the physico-chemical variable. "Axial gradients have often been called metabolic gradients", he said, "because differences in metabolism, or, more specifically, of oxidative metabolism, as indicated by various experimental methods, appear to be characteristic and conspicuous features of them." It will be necessary presently to examine the evidence on which this statement is based, but first of all a few theoretical remarks require attention. That gradients of various kinds exist within the developing embryo has long been known. The "law of developmental direction" (Jackson and Scammon) or of "cephalocaudal differential growth" (Calkins), which we have already discussed in relation to the growth of parts in the embryonic body, is simply, after all, a statement of the fact known in general to Aristotle, that the head end of an embryo develops quicker than the tail end. Again, as Minot showed, the cephalic somites develop before the more caudal ones. "The first parts to become morphologically visible", as Child puts it, "are the apical or anterior regions, and these are followed in sequence by the successively more posterior or basal parts." Again, there are dorso- ventral gradients. In those bilaterally symmetrical invertebrates which have a ventral nerve-cord (including most worms and arthropods) the ventral and median regions of the embryo at any given level of the body develop more or less in advance of the dorsal and lateral regions. On the other hand, in vertebrates, where the nerve-cord occupies a dorsal position, the dorso-ventral gradient runs the opposite way, and differentiation and growth proceed more rapidly in the median dorsal region than in the lateral and ventral regions. The antero-posterior gradient, however, is the same as in the invertebrates. For a recent discussion of the law of cephalocaudal differential growth Kingsbury's papers should be consulted. Again, the rule which has been named after F. M. Balfour, that the rate of cleavage in an embryonic region is inversely proportional to the amount of yolk which the cells in it contain, is associated with the gradient system of the egg. For the apicobasal gradient in the 584 ON INCREASE IN COMPLEXITY [pt. iii amphibian egg, for instance, leading from the protoplasm-rich rapidly dividing animal pole to the yolk-rich slowly dividing vegetal pole, has an enormous effect on the type of development which takes place. It is plain, too, that much more might be said about the relation of gradients to the various classes of eggs, alecithic, telolecithic, and centrolecithic. In 1905 Morgan advanced the hypothesis that the gradation of materials in the egg was a factor in establishing physiological polarity, and Boveri came to very similar conclusions about the Ascaris egg. Child made a step forward from these simple facts when he propounded the conception of primary protoplasmic gradients of which the morphological gradients were the obvious result, for he brought the subject to some extent nearer the point where physicochemical analysis could begin. Moreover, he was above criticism when he did not specify what sort of physico-chemical activity it was that was responsible for the gradient. But it is difficult to follow him when he concludes that the gradient is one of "oxidative metabolism", "oxidising power" or metabolic rate. Apart altogether from the fact that the experimental evidence will not carry this conclusion, there are theoretical difficulties involved in it. (It seems to have its roots, indeed, in that now abandoned idea, ^ which was once common among the followers of Jacques Loeb, that oxidation-processes and growth were very closely allied, even that the master reaction of the growth-phenomenon was an oxidation. If no physiologist now adopts this notion it is because so many researches have shown it to be false. The work of Crozier and his school on temperature characteristics might be mentioned, in which the [x value for growth in general and embryonic growth in particular practically never turns out to be 16,000. Murray's demonstration of the diametrically opposite course taken by growth-rate and metabolic rate, again, is an instance of the same thing. All the tendencies of recent years have been against any close identification of oxidative processes with growth. A little reflection is enough, moreover, to convince one that such an association is far from being a priori necessary. In the case of the embryo of the chick in the egg, for example, its increase in size could be represented by the curve a-b, in Fig. 95. At the beginning and at the end of the period x-j> the size of the embryo is of course given by the height of the curve above the abscissa, but to conclude SECT. 3] AND ORGANISATION 585 Absorption that the amount of solid absorbed by the embryo during that period was equal to the difference between the size at x and the size atj would be to assume that the efficiency of the embryo was 100 per cent., which is certainly not the case. As the diagram shows, what has actually happened could be represented in abstract form by a peaked curve superimposed on the growth-curve representing by its upward sweep the total quantity of soHd absorbed during the period in question and by its downward fall the total quantity of solid cataboHsed, or "oxidised", and excreted as carbon dioxide, uric acid, etc. The fact that the downward slope does not go down as far as the upward slope has come up is what makes growth possible. The process of growth, then, might be related to that of cataboHsm as rival, not as offspring, and instead of resulting from it might compete with it for the available solid substance. Or, on the other hand, the steeper the curve the higher the peaks might be above it, in which case metaboHc rate would be highest when growth-rate is highest. Everything depends on the relative magnitudes. The experimental fact that the metaboHc rate of the whole embryo is highest when the growth-rate is also highest reveals no simple causal nexus between them, and it is not theoretically correct to assume such a relation. Child has used various methods to ascertain the existence and distribution of his gradients: Fig. 95 (I (2 (3 (4 (5 (6 (7 Direct susceptibility. Indirect susceptibiUty. Reduction of potassium permanganate. Formation of indophenols. Observation of electrical potential. Estimation of carbon dioxide produced. Estimation of oxygen taken in. 586 ON INCREASE IN COMPLEXITY [pt. iii It is clear that the only methods capable of informing us whether gradients of metabolic rate are involved (cubic millimetres of oxygen used up per gram per hour or cubic millimetres of carbon dioxide given off per gram per hour) are direct estimations of these gases. Yet out of the hundred odd papers which make up the core of the literature on physiological gradients, not more than a dozen at the very outside are concerned with these fundamental measurements. In 1915, when Child's first two books were published, the only evidence available was due to Tashiro, who had, at Child's request, examined the behaviour of planarian worms in his microrespirometer, and had concluded that the pieces from the cephalic end gave off more carbon dioxide relatively than those from the caudal end, though the figures seem never to have been published. In view of the criticisms which Adam; Bayliss; and later Parker brought against Tashiro's apparatus, not much weight can be attached to these results. In 1 92 1 Robbins & Child obtained evidence from a study of regeneration in planarian worms that the larger amount of carbon dioxide was produced relatively at the head end, and in the following year Hyman & Galigher reported the same relationship to hold as regards oxygen for the oligochaete worms Lumbriculus inconstans and Nereis virens. Preliminary results on Corymorpha palma, a large tubularian hydroid, were given by Child & Hyman to the American Zoological Society in 1922 and 1923, and published in extenso by Child & Hyman in 1926. Hyman extended this to oxygen uptake of planarians in 1923. Her results were paralleled by figures for carbon dioxide production estimated by a colorimetric method in which the time taken by pieces of the stem to reach a definite acidity was measured. But no precautions were used to ensure that the acidity measured was due to carbon dioxide and not to other acids, for the method did not involve passing a stream of air through the water containing the stem under investigation. The oxygen determinations, on the other hand, were all done by the Winkler method, and were never checked by any differential manometer technique. That this is a serious deficiency is evident from the remarks of Shearer, who in a private communication says, "With Haldane's and Barcroft's apparatus I could not get any results which even begin to support Hyman's tables. I think that the Winkler method is useless where a lot of slime is discharged into the water (as is the SECT. 3] AND ORGANISATION 587 case with Planarians) ". This view finds many supporters among those famiHar with the Winkler method. Hyman and Child made an attempt to gauge the concentrations of glutathione at the different levels of the stem of tubularians, using a modification of the nitroprusside test, from which they concluded that glutathione gradients were present. Such a statement, 15 en X O 10 o - Heado \ V Tail X- — ^ >» "^. >» —X ■^ -^f— — X Days 4 5 6 7 8 9 10 Fig. 96. however, cannot be accepted in the absence of exact quantitative data, especially considering the unspecific character of the nitroprusside test applied to an intact animal. Apart from these rather unsatisfactory researches, nothing has been done by Child and his collaborators on the direct verification of their theory. In spite of this, they carried over the conception of metabolic gradients to the developing embryo without modification, and the question naturally arises, to what extent were they justified in doing so ? The position is, in a word, that it cannot yet be considered as proved that gradients of metabolic rate exist, much less that they accompany gradients of other entities, such as susceptibility and 588 ON INCREASE IN COMPLEXITY [pt. iii electric potential, even in the case of invertebrates — tubularians, planarians, annelids, etc. Are we then justified in asserting, direct evidence being absent, that a similar relation holds for vertebrates and embryos in general? Hyman's few figures for respiration of Fundulus eggs tell us nothing about the metabolic rate, for no weighings of embryos were made. As regards the embryo, it is most unfortunate that the only other piece of relevant evidence is contradictory. Shearer in 1923 made an investigation of the oxygen consumption of the head end and the tail end of chick embryos, using the Barcroft differential manometer. Fig. 96, taken from his paper, shows the results which he obtained. On the 4th day the head pieces took up more than three times as much oxygen per gram per hour as the tail pieces did, but by the loth day the two curves had almost come to coincide. Both of them fell, showing that the metabolic rate of the cells was declining with time in the usual way. Such a graph fits in very well with the morphological picture, for by the loth day the axiate pattern has long been established, and subsequent development mainly concerns growth in size. Shearer went on to investigate the action of acetone powders in order to see whether the higher metaboHc rate of the head pieces was dependent on structural conditions. The powders were made in the same way as acetone yeast preparations by first dehydrating the tissue in acetone, and subsequently desiccating completely. Such powders on being made into a thin emulsion with distilled water respire. The results were as follows: c.c. oxygen taken up by embryos of 6-7 days' development per amount of tissue containing 1-4 mgm. nitrogen. Exp. Head Tail 1 0-62 0-23 2 0-52 0-29 3 0-47 0-27 •This then aflforded a clear demonstration in favour of the identification of metabolic rate gradients with axial gradients, and gave evident support to Child's views. Unfortunately, in a later statement Shearer reported that he had not been able to repeat these findings, and that further experiments had very much modified the original conclusions. "I have since concluded", says Shearer in a private communication, "that I was dealing with the rate of cytolysis SECT. 3] AND ORGANISATION 589 undergone by the head and tail fragments. The younger the embryo the more readily the tissues disintegrated and cytolysed during an experiment. The heads containing the large watery brain vesicles cytolysed very quickly and gave a wholly abnormal respiratory rate, but the tail fragments, being composed of less deHcate material, did not undergo cytolysis so quickly. Whenever you get any cytolysis the oxygen consumption is of course greatly increased." Shearer afterwards extended the work with Barcroft microrespirometers to planarian head and tail fragments, and absolutely failed to get the results which should have been found on the Child theory. Here the position was compHcated by muscular movement which did not go on to the same extent in the head and tail fragments. The situation is therefore at present a deadlock, and we are at a standstill until further accurate work on the lines already laid down by Shearer is carried through. Practically no importance can be attached to the experiments which have been made with potassium permanganate and indophenol blue. Child & Hyman in 19 19 placed embryos in very dilute solutions (Af/ 10,000) of potassium permanganate, and described in all cases a gradient along the cephalocaudal axis with maximum activity, i.e. maximum reduction of the permanganate, at the anterior end. Similar work was afterwards done by Child (on Corella), by Galigher, and by Hyman. As a demonstration of contributory interest the permanganate method has its value, but it is far too uncertain biochemically to serve as the basis for the identification of axial with metaboUc gradients in embryos. As for the indophenol blue reaction, Child applied it to the development of the starfish egg, and observed that the apical third or half of the body of blastula and gastrula stages was always stained a deep blue before the blastopore region had become stained at all. In his 1924 book he stated that exactly similar gradients had been observed with methylene blue, reduction of this dye being faster at the cephalic than at the caudal end of the embryo. These observations cannot by any means permit of conclusions about gradients of metabolic rate. These criticisms of one of the main aspects of Child's theory are essentially the same as those made by Parker and by Loeb in his book on Regeneration. "The unit for the measurement of metabolism", said Loeb, acidly, "is the calorie, and the calories produced by an 590 ON INCREASE IN COMPLEXITY [pt. iii animal or by one of its segments are not measured by the time required to dissolve the animal or one of its segments in a solution of potassium cyanide." Nevertheless, the susceptibility method does show up the existence of gradients of something. The very numerous studies of Child; Hyman; and Bellamy on the susceptibiHty of different regions of embryos at different stages are quite sufficient to prove that. They will therefore be called in the remainder of this discussion axial or physiological gradients, and their nature will be left undefined. An example not tied up with any theory of the nature of the gradients is the staining gradient of the chick embryo found by McArthur to hold for many acidic and basic vital dyes. It must be admitted that throughout Child's treatment of the subject he confuses growth-rate with organisation- or differentiation-rate though the two are certainly not identical. In fact, it is often difficult to tell from Child's arguments whether he means growth-rate, differentiation-rate or metabolic rate, and it is not very helpful to lump them all in one as "level of physiological activity". But it is time to come to the facts obtained by the use of the susceptibiHty method. Child used two variations of the susceptibility method, the direct technique and the indirect technique. In the former case the resistance or susceptibility is determined directly by concentrations of toxic agent which kill the animals within a few hours. For a particular species a concentration must be determined which kills without acclimatisation, but which does not kill so rapidly that no differences between parts of the body are discernible. Child has used all kinds of toxic agents in his experiments, various cyanides, alcohol, ether, chloroform, chloretone, acetone-chloroform mixtures and other narcotics, X-rays, ultra-violet rays, ammonia, soda. He has obtained the best results with those substances which have a narcosis time and a killing time very close together, in other words, with those substances whose effects are not complicated with narcosis. The most favourable poisons, he found, were the cyanides. As Hogben has pointed out, it would have been far more favourable to the metabolic gradient theory if Child had confined his attention to the cyanides, which are known to have an inhibitory action upon some, though not upon all, tissue oxidations. For then it would not have been known that substances such as ether and chloroform had precisely the same effects, and the association between axial gradients and oxidation-rate would have been more convincing. SECT. 3] AND ORGANISATION 591 Since the death and disintegration of the different parts of the body usually follow a regular time sequence, Child found it possible to determine the time not merely of disintegration of the whole animal, but of various regions of the body. The method used as a rule was to examine the lot of animals at intervals of half-an-hour and then to record the condition of each individual. Arbitrary values having then been attached to the various stages of disintegration, curves can be constructed showing the rate at which the process has gone on. Fig. 97, taken from Child's 1915 book, shows such a curve constructed for specimens of a flatworm, Planaria dorotocephala, the curve ab representing the susceptibility of young and the curve cd representing that of old animals. The correlation between age and susceptibility should be noted. The indirect method involves the principle of acclimatisation. In general, according to Child, the ability of an animal to acclimatise itself to cyanide or other toxic agent varies with its metabolic rate, or rather its level of physiological gradient. The ability of parts of animals to ^^' ^^' become acclimatised also alters with the same variable, so that the indirect method affords a way of estimating gradients in the embryo. The relation between age and survival time in solutions of the concentration necessary for the indirect method is exactly the converse of what it was in the direct method, for here the younger animals with the higher metabolic rate live much longer than the older ones with the lower metabolic rate. The susceptibility method is obviously a very complicated one, and conclusions from experiments with it have to be drawn with caution. In lethal doses which are not concentrated enough to produce death within a short time after the beginning of the exposure, the regions of higher activity are always affected first, so that above Hourso 1 592 ON INCREASE IN COMPLEXITY [pt. hi a critical concentration susceptibility varies directly as the physiological activity, while below this concentration the reverse of this relation is seen, in that regions of higher activity recover and adjust themselves to the reagent more successfully than regions of lower activity. "In applying the susceptibility method to embryonic development, lethal concentrations may be used but not allowed to act long enough to produce death in the embryo, and in such cases they will, according to Child's interpretation, inhibit regions of higher activity to a more marked degree than regions of lower activity ; while on the other hand, in very low concentrations of the reagent such as to permit acclimatisation and recovery, the region of higher activity will be inhibited, according to Child's interpretation, less than regions of lower activity." A simple instance of the operation of Child's conceptions is the case of the eggs of some polychaete worms, Chaetopterus and Nereis, which, when placed unfertilised in lethal solutions of cyanide, exhibit a progressive dissolution from the anterior end. As development proceeds, the region of maximum susceptibility shifts round to the posterior region (where growth is most active) so that, when the larva is ready to metamorphose, the posterior region is the region which succumbs most readily to lethal and recovers most readily from sub-lethal concentrations of cyanide. Child's work on the eggs of the starfish, Asterias forbesii, brought out the fact that the susceptibility gradients in the unfertilised egg were connected in some way or other with the mode of attachment of the individual egg to the parent body in the ovary, obviously a very important point for the question of the origin of polarity and axial symmetry in the developing embryo. Wilson & Matthews showed that the region where the nucleus lay nearest the surface became in the starfish's tgg the apical or animal pole, and Child found that it was from that point on the surface of the egg that disintegration began. From the behaviour of nuclei which had been extruded during cytolysis in cyanide. Child concluded that the nuclear susceptibility gradient ran in the same direction as the cytoplasmic susceptibility gradient. The direction of the axis, he concluded, was determined by the eccentricity of the nucleus. Child found that during the earlier cleavage stages the general gradient was obscured to a great extent by individual gradients in the blastomeres and other incidental factors, and that "in the early spherical SECT. 3] AND ORGANISATION 593 blastula before movement begins the disintegration gradient is distinct, but the difficuky in identifying the animal pole and embryonic axis makes it impossible to demonstrate that the gradient coincides with the axis". But "in later free swimming stages of the blastula the direction of movement with the apical region, the animal pole, in advance, and before gastrulation the elongation of the embryo in the direction of the axis and the increasing thickness of the cellular layer toward the vegetative pole render orientation possible at a glance. In these stages the disintegration gradient is very distinct. It begins at the apical end and proceeds with a definite course along the embryonic axis, ending in the region of the vegetative pole where the gastrular invagination will occur". The susceptibility gradient in the gastrula of the starfish Child found to be very similar to that in the blastula, being greatest in the region of the apex and least in the region of the blastopore. Moreover, the gradient in the archenteron wall was exactly the same as that in the body-wall, the apical end of the endodermal invagination being the region of greatest susceptibility. After this time the gradients become less and less distinct until by the bipennaria larva stage they have faded away almost entirely. Hyman extended in 1 9 1 6 the observations of Child on polychaete eggs to those of a microdrilus oligochaete, Tubifex tubifex. She found that, in the stage when the embryo has begun to elongate, its posterior region was the most susceptible to cyanide, and the susceptibility decreased as one passed forwards. In later stages the head end became more susceptible, and finally exceeded greatly the tail end in susceptibility, so that at hatching it was much the most easily disintegrated region. Thus the posterior region of high physiological level which is characteristic of the adult annelid arises very early in development. After hatching, a worm placed in cyanide disintegrates first at the head end, later at the tail end, so that the two waves of disintegration reach and fuse at a point posterior to the middle of the worm's length. Figs. 98 and 99, taken from Hyman's paper, show the degeneration of embryos of different stages in cyanide. Child and his associates frequently correlated their susceptibility gradients with gradients of electric potential. Hyman & Bellamy in 1922 gave a full account of the work on this subject with a critical discussion, and at the same time reported their results for frog embryos. Hyde had previously found the heads of recently hatched N EI 38 594 ON INCREASE IN COMPLEXITY [pt. iii toad tadpoles to be galvanometrically negative to the heads, and this Hyman & Bellamy confirmed for the frog. "The idea is advanced", they said, "that differences of potential in organisms, particularly the permanent differences which exist along the main axes of animals, are due to differences in metabolic rate at different regions, the region of highest metabolic rate being the most negative in the external circuit, most positive in the internal circuit." In the tadpole, therefore, the highest physiological level appeared to lie towards the tail. Fig. 98. ^. Fig- 99 Hyman has published a series of papers on the susceptibility gradients of vertebrate embryos. The first of these she devoted to the teleost embryos {Fundulus heteroclitus (minnow), Ctenolabrus adspersus (cunner) and Gadus morrhua (cod)). She showed that, in addition to the primary gradient of the embryo which has its high level pole at the anterior or cephalic end, there were also other "secondary" gradients arising from regions of high susceptibility other than the anterior end. Certain organs, also, may have their own axiate pattern, notably the heart. The cod and cunner embryos were studied with cyanide in the usual manner, but the impermeable egg-membranes of the minnow made this impossible, so that ammonium hydroxide had to be used instead. In the early blastoderm stages of the cunner and the minnow, the central cells were observed to be the most susceptible, and from them disintegration proceeded to the periphery of the blastoderm. But in the case of the cod exactly the reverse relationship held true ; SECT. 3] AND ORGANISATION 595 the periphery of the blastoderm was more susceptible than the central part. In the later blastoderm stages of the cunner the region of high susceptibility is shifted posteriorly, and a certain area along the margin of the blastoderm succumbs very readily indeed to the toxic agent. This is exactly where the embryo is about to arise. The eggs of the minnow could not be examined at this stage. In the cod, the germ ring was always much more susceptible than the central part of the blastoderm, and at its circumference one region is more susceptible than the remainder. This is where the embryonic shield originates, so that the conditions in the cunner and the cod are now very similar. As the embryonic shield grows forward, its anterior margin is most susceptible, and disintegration extends posteriorly from this. Slightly later stages, when the embryo is visible in the centre of the embryonic shield and the germ ring has advanced more than half-way over the yolk, were not observable with certainty in the cunner and the minnow. But in the cod they were clear enough, and here toxic action obviously began at the anterior end of the embryo, spreading backwards towards the posterior margin of the shield. Still later, at the time of closure of the germ ring, disintegration gradients were observable with ease in all three species, and always the susceptibility was highest anteriorly, diminishing and spreading backwards. The eyes are not very susceptible, and do not degenerate until the wave has passed half-way back along the neural tube. After the germ ring has closed, a secondary region of high susceptibility appears at the posterior end of the embryo. From this point onwards, there is no change in the gradients; there is a powerful spreading backwards from the cephalic zone of high susceptibility and a slight spread forwards from the caudal zone, with no complicating factors. The minnow differs from the cunner in possessing the two zones from the very earUest stages, and at certain early points in development, the posterior zone is the more important of the two. But as development proceeds the posterior zone declines in susceptibility and the anterior one increases, especially after the arrival of the optic vesicles, which fall an exceedingly easy prey to the toxic agent. In very late stages in the minnow, a region of high susceptibility develops in the hind brain where the cerebellum is forming. By this time the tip of the tail has become free from the yolk and somewhat more susceptible again. The cod embryo behaves in 38-2 596 ON INCREASE IN COMPLEXITY [pt. iii much the same way as the other two in the later stages, always having two regions of high susceptibility, the anterior preceding before the closure of the germ ring and the posterior preceding afterwards. Minor variations in susceptibility of eyes, fore brain, etc., were noted. The somites in all three teleosts disintegrate from each end, but more from the anterior than from the posterior. Special observations were made on the heart gradients, which agreed in many particulars with subsequent work by the same author on the gradients of the embryonic heart of the chick, e.g. the venous end was the more susceptible, and the gradient decreased towards the arterial end. Hyman was able to draw several conclusions from this work important for pure embryology, such as that, in different teleost embryos, the amount of material contributed to embryo formation by the germ ring is variable, being very little in the cunner and considerable in the minnow. This reconciled the views of older workers, such as Morgan; Sumner; and Kopsch. These points, however, together with the fact that her results gave no support to the concrescence theory, are not so important for the present purpose as the delineation of the regions of high susceptibility for comparison with other embryos. The double gradient (anterior and posterior zones of high susceptibility) is also regarded by Child & Hyman as important for a comparison which they make between segments in segmental animals and separate individuals, suggesting that, whereas in annelids the posterior zone is permanent and never comes under the control of the anterior zone, in vertebrate embryos it eventually dies away, so that further segmentation ceases. This need not, however, detain us here. In the same paper as has already been mentioned, Hyman measured the rate of oxygen consumption of Fundulus eggs during their development. Unfortunately, the figures do not give us any information which would either support or weigh against Child's theory of metabolic gradients. Owing to the small size of the embryo, the determinations had to be expressed in relation to looo eggs (i.e. embryos + yolks), and, as we have no idea how the wet and dry weights of the eggs or the embryos were varying during this period, we cannot calculate the metabolic rate. This work will be discussed in detail in the section on the respiration of the embryo. Hyman also made a summary of the teratological results which had SECT. 3] AND ORGANISATION 597 been obtained by various workers on Fundulus, and concluded that in all cases the malformations produced most easily were those of the fore brain, the head in general, the sense organs, especially the eyes, the heart, the circulatory system, and the tail. These results obviously fitted in very well with those appearing from the use of the direct susceptibility method of Child, and this outcome of the physiological gradient conception is perhaps one of its most attractive aspects, for no other point of view serves to account for so many of the facts of teratology. It was long ago pointed out by Dareste that no relation seemed to exist between the application of a certain physiological or physical condition and the resulting teratological modification. Thus Herbst's "lithium larvae" and Stockard's "magnesium embryos" have been shown to be obtainable with a great variety of agents. There is, we may say, practically no specificity in teratological action. "Any type of abnormality", as Bellamy puts it, "may be produced under the influence of any inhibiting agent by controlling the concentration or intensity of action, the length of exposure, and the stage, i.e. physiological condition, of the Ggg or embryo or parts of the egg or embryo when exposed." Since the differences then do not reside in the teratological agents employed, they must do so in the embryo itself — an admirably Kantian conclusion, which can only be explained on some basis which maps out the embryo into a logical system. The only basis we have is the conception of physiological gradients. It is difficult also to imagine any other view which could explain such instances as the production of the usual terata by fertilising eggs with foreign or injured spermatozoa or treatment of the eggs before fertilisation as in the experiments of Gee. Full discussions of the teratological literature and interpretations of it from this physico-chemical standpoint will be found for the fish embryo in the paper of Newman, for the chick in the paper of Hyman, and for the frog in the paper of Bellamy. It is very noteworthy that new teratological modifications not before obtained have been predicted on the basis of physiological gradients, and have afterwards been verified. Hyman next studied the gradients during the development of the brook lamprey, Entosphenus appendix, an organism of considerable interest, in view of the fact that, like the amphibia, the cyclostomes are one of the three vertebrate groups which develop by holoblastic unequal cleavage. Alcohol and acetic acid were used as the reagents ,.-Of|j^ ■■^afet Fig. lOO. Physiological gradients in the egg of the lamprey (Hyman) . 1-4. Unfertilised egg. 5-8. Four-cell stage. 9-12. Eight-cell stage. 13-18. About thirty-two-cell stage, 16 hours. 19-22. Morula, 20 hours. Fig. loi. Physiological gradients in the egg of the lamprey (Hyman). 23-26. Early blastula, 24 hours. 27-31. Late blastula, 40 hours. 32-35. Beginning of gastrulation, 48 hours. 36-39. Gastrula, 60 hours. 40-43. Late gastrula, 70 hours. 44-46. Late gastrula, neural groove about to appear, 75 hours. From no. 36 onwards, the anterior end is to the right. 6oo ON INCREASE IN COMPLEXITY [pt. m for producing differential death. An interesting point was that the embryos were first stained with neutral red, and the change in tint of this indicator noted as soon as the killing solutions were poured on them, obviously showing that the differences in killing time of different parts were not entirely due to differences of permeability. Cannon and Huxley were at one time inclined to attribute most of the results obtained by the direct susceptibility method to such differences. But in any case differences of cell-membrane permeability would be included in the variables which might be changing along the physiological gradient. The disintegration gradient of the unfertilised lamprey egg was found to be a perfect example of a simple primary gradient, the degeneration beginning at the animal pole and spreading regularly to the vegetative pole. At the 4-cell stage the disintegration begins at the animal tips of the four blastomeres, and passes backwards to meet a slight secondary zone of high susceptibility at the vegetative ends. These changes are seen in Figs. 100 and loi, taken from Hyman's paper. In the 8-, 12- and i6-cell stages, the degeneration begins in the micromeres at the animal pole, then passes on through the macromeres. In later stages the disintegration constantly begins at the animal pole, but in addition isolated cells or groups of cells are to be seen in either animal or vegetal hemisphere which disintegrate in advance of the region in which they are situated. Hyman supposed that these were taken by the killing solution in the act of cleavage, and were thus more susceptible than their neighbours. The secondary region at the vegetal pole is now disappearing for good. The early blastula stages show the usual single gradient, but an important change occurs in the later blastulae, namely, that the spread takes place more rapidly along one surface of the egg than the others. This foreshadows the differentiation of that surface as the dorsal surface. In very late blastulae, there is a small zone of susceptibility near the vegetal pole, which foreshadows the gastrular invagination. The gastrula stages, as shown in the figures, are characterised by disintegration beginning at the anterior end of the embryo and proceeding backwards, but first dorsally and then ventrally to meet the spread from a secondary zone originating around the blastopore. These conditions continue unchanged during the formation of the neural groove and the neural tube, though in the late stages of the latter there is a slight double SECT. 3] AND ORGANISATION 601 gradient in it. No further changes occur during elongation and hatching. The gradients in the chick embryo were also studied by Hyman, using potassium cyanide, and ammonium and sodium hydroxides, sometimes preceded by staining with neutral red. The earliest stages were very difficult to deal with, but some evidence was obtained of an antero-posterior gradient in the central opaque area of the germinal disc at 7 hours' incubation. This was quite certainly demonstrable, however, at the typical primitive streak stage, and the stage of the head process. The medullary plate stage marked the beginning of the double gradient (see Fig. 102), two regions of high susceptibility being present, one at the anterior end of the primitive streak and the other at the anterior end of the medullary plate. When the first somites appear the same zones are seen, the former spreading backwards and forwards along the embryonic axis, the latter backwards only. As the neural folds close, they present a region of high susceptibiUty, but this soon disappears, and the embryo reverts to the simple double antero-posterior gradient system. This holds good up to the 8-somite stage; from the 9th onwards the rapidly increasing susceptibility of the optic cups is noticeable. At the i2-somite stage the optic zone has died away and there is a new one of high susceptibility in the hind brain, foreshadowing the turning of the head, but this also disappears by the 3rd day of development. Summing up the results, one may say that the general picture is one of an antero-posterior gradient complicated from time to time by the appearance of zones of high susceptibility at different points along the embryonic axis. In vertebrate embryos in general, it would seem that the formation of these two regions of high susceptibility is the regular mode of development. The two centres are always located in the same position with respect to the future embryo, one at the anterior end of the antero-posterior axis and one in the axis at a more or less posterior point. This posterior centre is the dorsal lip of the blastopore in cyclostome and amphibian embryos, the posterior end of the embryonic axis in teleostean fishes, and the primitive knot, subsequently the tail bud, in the chick. "This posterior centre", says Hyman, "is Hke a growing point which, passing backwards, deposits the trunk of the embryo anterior to it." The presence of two centres of activity was long ago recognised in frog and rabbit embryos p;,r ,no Phvsioloeical grad ents in the chick embryo (Hyman) tig. I02. rnysioiogic^i gi'i" A/f<.rq,.ll^rv n ate sta£ e 4-6. Head-process stage. ?■ romi.fs?ie;t6i7°FTv:rm!,'ho-;^g'h?rgh^^^^^^ fold stage. i4-i7- One-somite stage somite stage. 26-; . of fusion of neural folds. i-Q. Primitive streak " ly ' 18-20. Three-somite stage. 21-25. I the cnicK emuiyu i^ixyx^ciw;. . -J. ' PcrWrK-nral ..age. 4^. Hjad-process stage. 7-.0. MeduUary (.a.e staje. ^. ,-.3. Early neu„^^ PT.ra,sECT.3] ON INCREASE IN COMPLEXITY 603 by Assheton, who referred to them as primary and secondary centres of cell-proHferation, and this point of view has been adopted by many embryologists, e.g. Eccleshymer; Adelmann; and Kingsbury. In addition to these researches on embryos, Child has studied the gradients during the development of the sea-urchin egg {Arbacia punctulata), those of polychates {Nereis, Chaetopterus and Arenicola) and those of an ascidian [Corella willmeriana) . In these cases, the primary simple apicobasal gradient of the fertilised egg-cell was succeeded by a double gradient resulting from the appearance of a zone of high susceptibility at the posterior end of the embryo. Mention of the work of Bellamy on the amphibian egg brings us back to our starting-point, namely, the recent investigations on the organiser and the mechanics of amphibian development. There is no need to give a detailed description of Bellamy's results on the frog embryo, for they resembled in many ways those of Hyman on the brook lamprey. But they may be briefly summarised for the purpose of comparing them with the work of Spemann and his school. Bellamy found that in the unfertilised amphibian zgg the beginnings of polarity were to be found in the position in the ovary. He observed by injections and by actual observation of blood-flow that the bloodvessels to the eggs in the oogonia pass arterially over the pigmented part of the egg and venously over the unpigmented part. It is more than probable that the first polarity of the egg arises because the animal pole is that point on the surface of the egg which happens to be most well supplied with a capillary network, not that which happens to be attached to the ovary by the pedicle. The initial physiological gradient, therefore, would seem to be a matter of position in the ovary. It may be mentioned here that very similar conclusions were come to by Lillie for Chaetopterus and Sternapsis eggs, by Child for Phialidium (hydromedusa) eggs, and by Boveri and Jenkinson for Strongylocentrotus eggs. The early stages of development in the frog's egg are very resistant to toxic agents. But it was possible to show that the fertilised but undivided egg began to disintegrate at the animal pole and the degeneration passed downwards and outwards with a special bias towards the grey crescent. Much the same state of affairs was seen in the 4-cell stage, but in the morula stages there are two zones of high susceptibility, the second one appearing just above the grey crescent, and contributing to the general spread down 6 04 ON INCREASE IN COMPLEXITY [PT. Ill wards from the animal pole. Eggs in an early gastrula stage always disintegrated first at the dorsal Up region and shortly afterwards in the same meridian about 120 to 130° above the blastopore. From the upper point at the animal pole the wave spreads downwards, and meets the disintegrated area of the dorsal Up, which has spread apically and now includes the lateral lips, after which all the pigmented cells become gradually involved, though the cells at the vegetal pole retain their structure with their yolk long after the rest of the egg has died. Lateron, when elongation has begun, the dorsal lip now takes up the posterior position, and the double gradient still persists, disintegration beginning from both ends, from the apical point at the anterior end and from the dorsal lip at the posterior end. This description applies to most of the later Fig. 103. stages, including the time of opening and closing of the neural folds, the appearance of ventral suckers, etc. Eventually, with the differentiation of various organs, local susceptibility differences begin to appear, and the tail bud, the optic vesicles, the nasal pits, and other rapidly growing regions show much susceptibility. (See Fig. 103.) On the basis of these fundamental results, Bellamy was able to make and verify teratological predictions, and to control experimentally or modify development, e.g. the gastrular angle and the cleavage ratio (the ratio between the sizes of animal and vegetal pole cells). He was criticised by Cannon, whose main objection was that the effects of the toxic agents were not uniform at the different stages, but that the individual differences between eggs were so large as to invalidate Bellamy's conclusions. When Cannon did succeed in getting a lot of eggs to behave in the same way at the same time, he found results quite at variance with Bellamy's, e.g. the ventral, not the dorsal, region of neural tube stages was the more highly susceptible. These criticisms were replied to in detail by Bellamy & Child, who successfully rebutted them, and whose SECT. 3] AND ORGANISATION 605 paper should be consulted for further details. Other criticisms of the general theory of physiological gradients have been made by Wilson; Kingsbury; Lund; Allen; and others, but they do not affect the main conclusions which have been described. The significance of Huxley's remarks in 1924, which have already been quoted, can now be better appreciated. At the time of chemodifferentiation, the various irreversibly determined regions differ from each other by the presence not only of qualitatively different substances but also by the presence of varying concentrations of the same substance, according to the conception of axial gradients. "If it is asked", said Huxley, "how we can imagine the process as originating, the answer must, I think, follow some such lines as these. During gastrulation every portion of the embryo has a definite relation to the system of axial gradients. The two main gradients extend both on the surface and internally and together constitute a three-dimensional system of gradient co-ordinates. Every portion of the embryo, therefore, has its own rate of activity corresponding to its position in the existing co-ordinate system, and its own characteristic proportions of yolk, glycogen, cytoplasm, etc., depending on the previous effects of the apicobasal gradient during the growth of the egg. When the organiser in the dorsal lip exerts its admittedly as yet unexplained though not unparalleled action of initiating differentiation, every region of the embryo is in a different condition from every other. The substrate is different from place to place, the result, therefore, also differs." The position of a given point in the embryo on the physiological three-dimensional graph is thus of greater importance than the proportions of primary materials which it contains, and, further, the relative velocities of processes going on there are more important than the actual amount of substances of different kinds that happen to be present there. The substances which can be distinguished in the unsegmented ovum of a vertebrate are thus merely raw materials, and the organising influences are to a large extent expressible in terms of gradients of activity. The first of these is probably determined before the egg is laid at all, the second arises from the action of agencies external to the egg approximately at fertilisation. We are thus left with the conception of parts of the embryo as pacemakers of growth and differentiation relatively to the rest: though in what exactly their influence consists we do not as yet know. "The first step in organisa 6o6 ON INCREASE IN COMPLEXITY [pt. m tion and in embryonic development", says Child, "results from the establishment in one way or another of some region or portion of this protoplasmic reaction-system as a region of higher rate of dynamic activity. This region dominates development, becomes the apical or head region and determines the axial gradient or gradients which constitute the dynamic basis of polarity and of individuation." 3-9. Organised and Unorganised Growth An idea which has much importance for the study of the interrelations between growth and differentiation was contained in a paper by Faris on the pigmentation of Amblystoma embryos. The details of his investigations into the ontogenesis of this pigment will be referred to again in the section on pigments; here I am concerned to refer to his distinction between "proliferation metabolism" and "differentiation metabolism". He observed that, during the development of the myotomes of Amblystoma embryos, the pigment accumulated in the cells proportionally to differentiation and not to growth. He therefore suggested that it could be regarded as an index of the difference in nature between the type of metabolism associated with growth and that associated with differentiation, admitting, of course, that the two processes were only completely separable in the abstract. "Proliferation", according to Faris, "must lack the wear-and-tear processes that are characteristic of differentiation and for that reason it lacks the function of pigment production." It must be admitted that these concepts are vague enough, and they rest on an unsatisfactory, because unquantitative, basis. But they have a real interest, in view of two lines of recent thought : firstly, the distinction made by Murray between the groups of processes in embryonic development according to velocity at different times, and secondly, the numerous papers of Warburg and his school (see Section 4-20), which introduce into these problems the concepts of organised and unorganised growth metabolism. These investigations will be referred to in detail in the next section. If it should turn out that Faris's conception of two types of metabolism found a quantitative basis in the data of Warburg and his collaborators, an interesting and quite important avenue would be open for further investigation. The notion of a distinction between organised and unorganised growth had not, however, its sole origin in Germany, for Byerly had also come to it from very different ground. Setting out SECT. 3] AND ORGANISATION 607 from the idea originally suggested by Jordan that haemopoiesis was associated with lack of oxygen or accumulation of carbon dioxide, he allowed chick embryos to develop for 24 hours or so, and then, breaking off the shell covering the air-space, immersed the whole egg in water-glass solution so that the respiratory exchange was quite stopped, but incubation allowed to proceed till 96 hours. He then examined the suffocation effects so produced. These were (i) normal body-form only at the anterior end, (2) no allantois, (3) extraordinarily large blood-vessels and anomalous sinuses, (4) constantly recurring fatty necrosis of tissues. Since the circulation of blood had stopped in these embryos, the end-products of metabolism were accumulating in their cells. The enormous amounts of blood found obviously suggested an unusual haemopoietic activity. But the heart, being unable to beat in the toxic anoxaemic blood, allows it to accumulate in the vessels, and, as more blood-cells are continually being formed, sinuses develop. Cessation of the circulation having led to a struggle for existence between the various parts of the body, some may find it possible to live on the rest, and unorganised "anarchistic" or unregulated growth may occur. This is what Byerly actually found in the suffocated embryos, as regards the formation of blood, btit he did not bring forward any evidence that an unusual quantity of lactic acid was produced in the suffocated embryos, as would have been the case on Warburg's view if one tissue had taken on an anaerobic Hfe and was growing at the expense of the others. Holmes was later unable to repeat these observations of Byerly's on the chick embryo, but the number of her experiments was insufficient to negative definitely Byerly's conclusions. In his second paper, he continued the study of "dead" embryos, which he observed to show three types of behaviour. One class remained on the surface of the yolk and showed anarchistic growth, one absorbed liquid to form a bladder-like vesicle on the surface of the yolk, and one sank beneath its surface. Beha\dour of the tissues of embryos of the first class varied according to the age of the embryo at the time when the heart was made to stop beating. If cardiac failure occurred after 4 or 5 days' incubation, only the liver cells and certain of the blood-cells were still proliferating at the end of a week's anaerobiosis. But if the failure took place at 72 hours' incubation or less, then, in addition to liver and blood growth, the 6o8 ON INCREASE IN COMPLEXITY [pt. iii nervous tissue went on proliferating for nearly a week. In such embryos the mesenchyme cells seemed almost unaffected by the suffocation, and continued to divide, but took on the histological characteristics of blood-cells. The important point about the picture in anaerobic embryos was that each cell became a relatively free unit, and all correlation of growth as well as all further differentiation ceased with the circulation. Only unorganised cell-life was possible. No consideration of organisation in growth could omit the subject of mitogenetic rays, from which it appears that radiant energy of definite wave-length is given off by cells in mitosis as an excitant to adjoining cells not in mitosis. For detailed information on this subject, the book of Gurwitsch & Gurwitsch and the memoir of Borodin should be referred to. Genuine uncertainty still exists, however, about these rays, and a whole literature is growing up, partly consisting of reports of workers who confirm Gurwitsch's results, and partly of the reports of those who do not. Anikin has already attempted to relate the activity of these " mitogenetische Strahlen" to organiser phenomena, and Sorin & Kisljak-Statkewitch have examined all the parts of the hen's egg, using onion roots as detectors for the rays. Negative results were obtained with albumen from the 2nd day of development, the vegetal pole of the yolk throughout incubation, the white and yolk of infertile eggs, even if incubated, the "Brei" of germinal spots 36 hours old, cerebro-spinal fluid and brain tissue from 5-day embryos, the amniotic liquid and certain other parts. On the other hand, positive results were always obtained with the substance immediately under the germinal spot up to the 6th day of development and with the blood. Karpass & Lanschina also find peptic and tryptic digests of egg-yolk to be powerful sources of mitogenetic rays. 3-10. Chemical Embryology and Genetics The relations between physiological and chemical embryology and genetics, on the other hand, afford a more solid basis for discussion. It has frequently been found that the behaviour of organisms of known genetic constitution during their embryonic period affords a means of marking out the points in ontogenesis at which genetic factors come into play*. The simplest case of this kind is the question

Some embryological factors seem to be determined entirely by the maternal organism (Toyama for pigmentation of silkworm eggs, Diver, Boycott & Garstang for dextrality and sinistrality of Limnaea eggs) . SECT. 3] AND ORGANISATION 609 of embryonic mortality, which will be mentioned again in Section 18, and which provides a striking temporal field for the display of genetic characteristics. To take only one example, Dunn & Landauer studied the relative embryonic mortality of a variety of chick called the "creeper". Cutler had first reported that, in "creeper" fowls, the leg and wing bones were shorter and thicker than in normal fowls, and that creepers never bred true, but usually produced {a) normal chicks, {b) creeper chicks and (c) chicks with extreme leg defects. The case was thus analogous to that of the yellow mouse, which is always heterozygous, because the homozygous yellow embryos die early in development. The relative embryonic mortalities were found to be as follows : _^ Percentage of total embryos incubated. Dead in shell Hatched alive 27-3 66-2 Creeper male x creeper female Creeper male x normal female 1-6 days 45-5 4-2 7-13 days 15-2 16-9 14-21 days I2-I 12-7 These results lead naturally to the assumption that the high early mortality of the creeper x creeper matings was due to the death of homozygous creeper embryos early in development. The creeper variation would thus seem to be due to a single dominant gene which is lethal in the homozygous condition. But the important point for this discussion is that, if the point of action of a gene can be found to occur at a definite point in development, a new outlook in genetics becomes possible, for what we know to be taking place in the physiological and chemical activity of the embryo at that period may tell us a good deal about the gene itself. Much thought has been given to these questions in recent years. Danforth, for instance, asked the question whether genes interact with one another during embryonic development to produce structural and functional characters, or whether they each exert their influence separately, thus Experiments with mice led him to regard the latter view as the more probable. The three known effects of the Y-gene occur N EI 39 6io ON INCREASE IN COMPLEXITY [pt. in at 6, 25 and 90 days after fertilisation of the egg, and a complex which modifies one of these effects may have no effect on the other two, e.g. if colour is replaced by albinism, yellow is entirely suppressed but senile adiposity is not influenced. Plunkett made a systematic analysis of the way in which genetic factors and environment interacted to produce bristles during the embryonic development of Drosophila. These bristles can easily be counted, and so afford a quantitative variable. What actually happens finally as regards bristles depends on (i) genetic factors, e.g. missing-bristle genes, extra-bristle genes, etc., (2) the temperature during the developmental period, (3) other environmental conditions such as nutrition, (4) differences of internal environment and (5) random internal variations. The last two of these can be avoided by suitable methods. Rise of temperature, Plunkett found, tends to suppress bristle formation. It must act, then, either by decreasing some bristle-forming reaction or by increasing some bristle-inhibiting reaction. If the temperature characteristic, he argued, of this reaction is less than development as a whole, the former must hold, but if it is greater, then the latter must hold. As the evidence came out, it was in favour of the former theory, for the critical thermal increment was 36,400, a value often found for heat destruction of enzymes, although for development as a whole it was 27,800 from 14 to 17°, 17,100 from 17 to 25° and 9000 from 25 to 30°. By various calculations, Plunkett showed that the reaction in question began very early in the larval stage, and took place in three steps: [a) production of R (a destructive agent for the enzyme) from some protoplasmic component, {b) destruction of B (enzyme from the gene) by R and [c] the formation of the bristles catalysed by B. A bristle-reducing gene might therefore throw its weight into the catalysis of the formation of R. Plunkett suggested that all genes act by differential acceleration of enzyme actions during embryonic development. Similar work on temperature characteristics of gene action has been done by Nadler; and Gowen, after 'K-ra.ying Drosophila at various stages of development, reported that the abnormalities produced, being strictly confined to certain groups of cells, showed that the gene had been affected prior to any action, and while it was still, as it were, lying latent in the nucleus. Huxley & Ford and Ford & Huxley have undertaken interesting studies on the eye-pigment of Gammarus chevreuxi. Allen & Sexton SECT. 3] AND ORGANISATION 611 had found that the red colour of the eyes in this organism darkened with development almost to black, and, by working with an arbitrary scale of colours, Huxley & Ford were enabled to make a quantitative examination of the rates of action of the various genes which influence eye-colour. Fig. 104, taken from their paper, demonstrates diagrammatically the relationships found. The blackening appears to be due to the deposition of melanin; in some cases this occurs very rapidly, in others more slowly. Thus the steepest curve in the diagram represents the dominant black-eye type which is black at hatching. Embryos of this type, however, which have not completed half their incubation, have no colour in their eyes, but soon they Dominant Black Eije. Below thj3 all are Recessive Red Moat Rapid Djrieninq Me^n Rapid Daitieninq Mean Slow Daritenm^ Absence of DdHteniiu) Fig. 104. become pale pink and later scarlet. Just before the end of embryonic life, the eyes darken, until at about the time of extrusion from the brood-pouch they are quite black. It has taken 10 days from fertilisation at 20° to bring this about. Other genetic types, however, hatch with red eyes, and only much later in life approach to blackness, as the diagram illustrates for various varieties. Thus, a definite relation was found to exist between Mendelian genes and rate of a chemical process. Morgan, in 1923, discussed such questions as these. "It is to be hoped", he said, "that in time the combined attack on the problem 6i2 ON INCREASE IN COMPLEXITY [pt. iii of development by genetics and experimental embryology and especially by chemistry may lead to the discovery of the physiological action of the genes, but for the present we may confess ignorance." A great step forward in this direction was taken by Goldschmidtt in his important book on physiological genetics. He assumes that genes are primarily of the nature of enzymes or substances which can excite the action of enzymes. Working on the caterpillars of Lymantria dispar he found that by crossing the European and Japanese races, intersexes could be obtained, and he was able to identify their various grades between the two poles of maleness and femaleness with the time-process of development, i.e. with critical points earlier or later in ontogeny. His conclusions were, firstly that the velocity of the sex-determining (and all morphogenetic) reactions was proportional to the quantity of the genes present, secondly that all the morphogenetic reactions go on side by side, the most rapid one controlling development and, thirdly, that the morphogenetic reactions involve determination-hormones brought into being by gene action. In other words, the genes evoke the organisers, which evoke the morphogenesis. It is obvious that the co-operation of the chemist and the geneticist in the investigation of embryonic development will be very fruitful in the future. With the work of Onslow on the chemistry of coat-colour and of Brink and his collaborators on the waxy gene in maize and its chemical effects, such co-operation may already be said to have begun. Brink & Abegg, in an important passage, point out that, though at present there is little likelihood of valuable results emerging from further chemicafstudy of nuclear material, i.e. the genes themselves, there is every chance of success in the investigation of the chemical field of action of the genes. The manner in which these units function in ontogeny plainly offers a prodigious field for the future work of chemical embryologists. Interesting reviews of this subject are those of F. R. Lillie and of T. H. Morgan. This section may fitly be concluded with the words of Sir W. B. Hardy: "Let us consider the egg as a physical system. Its potentialities are prodigious and one's first impulse is to expect that such vast potentialities would find expression in complexity of structure. But what do we find? The substance is clouded with particles, but these can be centrifuged away leaving it optically structureless but still capable of development. . . . On the surface of the egg there is SECT. 3] AND ORGANISATION 613 a fine membrane, below it fluid of high viscosity, next fluid of relatively low viscosity, and within this the nucleus, which in the resting stage is simply a bag of fluid enclosed in a delicate membrane. How shall sources and sinks of energy be maintained in a fluid composed of 80 per cent, of water? They are undoubtedly there, for the egg is a going concern, taking in oxygen and maintaining itself by expenditure of energy. . . . The egg's simplicity is not that of a machine or a crystal, but that of a nebula. Gathered into it are units relatively simple but capable by their combinations of forming a vast number of dynamical systems into which they will fall as the distribution of energy varies.'* END OF VOLUME I CAMBRIDGE: PRINTED BY W. LEWIS, M.A., AT THE UNIVERSITY PRESS

Cite this page: Hill, M.A. (2020, August 9) Embryology Book - Chemical embryology 1 (1900) 3. Retrieved from

What Links Here?
© Dr Mark Hill 2020, UNSW Embryology ISBN: 978 0 7334 2609 4 - UNSW CRICOS Provider Code No. 00098G