Talk:Paper - Normal and abnormal development of human embryos

Normal and Abnormal Development of Human Embryos: First Report of the Analysis of 1,213 Intact Embryos

HIDEO NISHIMURA, KIICHI TAKANO, TAKASHI TANIMURA, AND MINEO YASUDA

Department of Anatomy, Faculty of Medicine, Kyoto University, Kyoto, Japan

ABSTRACT This survey of the development of a large number of virtually unselected human embryos from healthy women allowed us to establish more reliable standards of normal development with respect to crown-rump length, body weight, and external form than those usually cited. Remarkable variation was noted with regard to the relation between clinical age and these attributes and thus it is concluded that crown-rump length and body weight are more reliable indicators of the general state of development than clinically established age.

Dead embryos occurred with a signiﬁcantly greater frequency in women with a history of genital bleeding during pregnancy than in those without. Externally malformed embryos were found with increasing frequency with advancing developmental stage, with the maximum figure of 3.92% at horizons 19-23. Malformations observed included external defects such as exencephaly, cyclopia, myeloschisis, cleft lip, and several limb malformations. It is of considerable interest that the incidence of most

of these defects was far higher than that observed in newborn infants.

Reliable information on human embryonic development, both normal and abnormal, is meager. Many of the data available at present are not entirely reliable, because they were derived from a small number of specimens obtained from spontaneous abortion or from therapeutic abortion of mothers with pathological conditions.

Revision of the Japanese Eugenio Protection Law in 1952 allowed qualiﬁed gynecologists to terminate pregnancy for sociomedical reasons, requiring only the request of both the mother and her spouse. This revision was made originally for the purpose of relieving a severe food crisis occurring at the time in Japan. It resulted in a large number of recorded induced abortions, about 1.1 million per year since 1953, and caused an abrupt fall in the birth rate (Aoki, ’62). The ofﬁcial statistics show that more than 90% of induced abortions were performed in the second and third months of pregnancy (Section of Statistics in The Secretariat of Ministry of Health and Welfare, Japan, ’63). Although such a situation would seem to provide an exceptional opportunity for obtaining a large number of human embryos, large—scale studies were not immediately attempted. This was primarily because the operation was usually done by dilatation and curettage, a procedure that severely damages embryos in most cases. However, we noticed that there were some occasions when intact embryos were obtained from operations by skilled gynecologists. This led to the establishment of a large team of selected gynecologists willing to provide us with their better specimens; an arrangement that allowed us to obtain standardized data on normal and abnormal human development during the stages of organegenesis, based on specimens derived from healthy pregnancies. The results presented here are based on studies of 1,213 intact specimens obtained from 1961 to 1966.

MATERIALS AND METHODS

With the cooperation of 970 gynecologists, practicing for the most part in small regional clinics, embryos were obtained during the second and third months of pregnancy. Pregnancy was interrupted by dilatation and curettage for sociomedical reasons. Cases terminated for strictly med 1 This work was supported by grants from the National Institutes of Health, Public Health Service, U.S.A. (HD 01401), the Association for the Aid of Crippled Children, New York, the Ministry of Educa Eoné Japan, and the Fujiwara Memorial Foundation, yo 0.

ical reasons, such as threatened or beginning abortion or chronic maternal diseases, were excluded from this report. Therefore all specimens were obtained from women with a clinically healthy course of pregnancy. The specimens were carefully Washed with saline to remove maternal blood coagula and ﬁxed in Bouin’s ﬂuid at the clinic where removed. On the next day they were transferred to 10% formalin solution for storage.

The 1,213 undamaged specimens, excluding several twin cases, were weighed and their crown-rump length measured after storage in formalin solution for 3-6 months. Careful examination was made under the binocular stereomicroscope to determine the developmental horizon, using the Streeter criteria (’42, ’45, "48, ’51), and to observe malformations or indication of intrauterine death, using criteria developed in this laboratory (Pearson et al., ’68). For each case the gynecologists completed a form that included questions concerning the existence of familial defects, parents’ age, and mothers reproductive history such as parity, menstrual cycle, and the occurrence of various diseases or medical treatments. The information in the completed forms is briefly summarized as follows: the parents lived mostly in urban districts in the central area of Honshu Island at about 33-37” latitude; most were judged by occupation to be at the middle socioeconomic rank; mean maternal age and parity were 30.10 3 0.18 years and 1.70 i 0.037, respectively; and

frequency of consanguineous marriage was 2.31%.

RESULTS AND DISCUSSION

1. Normal development

To establish the standard for normal development 675 specimens were singled out using the following criteria: mothers with menstrual cycles of 27-32 days, which were regular within a range of 2 days, and no history of genital bleeding during the terminated pregnancy; and the specimens showed no external malformations. Although there is a possibility that some of the externally normal specimens had visceral anomalies their number was probably low. The relations between paired variables including ovulation age, developmental horizon, crown-rump length, and body weight were examined and are plotted in ﬁgures 1-6. The estimation of ovulation age was made according to Knaus-Ogino’s method, i.e., by subtracting 14 days from the menstrual age.

a. Crown—rump length in relation to age. The results are shown in ﬁgure 1 where

the mean curves determined by other investigators are compared with that derived from the present series. Most striking is the fact that large age variations existed for length. For example, 10-mm embryos varied in age from 26-50 days and 20-mm ones from 42-58 days. The coefficients of variation are 13.40 and 9.84, respectively. Compared with the curves by Streeter (’42, ’45, ’48, ’51), Patten (’53), Witschi (’56), Arey (’65), and Iffy et al. (’67) our mean line is inclined more horizontally. This indicates that the mean length for any estimated ovulation age in our series is greater than other standards at early stages (3-4 weeks) but falls behind later (7-8 weeks). Thus in our series mean length increases more gradually with age. Streeter’s line differs most strikingly from ours in this regard but differences from the data obtained by some other studies not included here are also conspicuous. For example, the average length on days 3048 based on Takakusu and Ban’s (’63) data is generally about 3 mm shorter than our corresponding ﬁgures, whﬂe the average length at 4-8 Weeks obtained by Shirnarnura (’57) is 1-4 mm longer than ours.

b. Body weight in relation to age. The ﬁndings (ﬁg. 2) indicate that variation in body weight was also high for any given age, particularly from 40-60 days. Witschi (’56) presented a standardized table on body weight of embryos with only ﬁve ﬁgures, namely, 20 mg for 35-37 days, 50 mg for 38 days, 1000 mg for 50 days, 1500 mg for 56 days, and 1500-2000 mg for 50--70 days. Our mean line is more horizontal compared with the line plotted from his data. The mean body weight obtained by Shimamura (’57) for ages 5-7 weeks is generally heavier than ours. Nakamura et al. ('55) presented the following linear regression between age and body weight of embryos: = 8.6434 log Y + 65.6130 (X = menstrual age in days, Y = body

Are)! ('65) lffy of al. ('67)

Patten ('53) Streeter ( '42- '5!)

Witschi ( "56 )

20

5

CROWN-RUMP LENGTH IN MM '6

30

Nishimuro am. ('68)

283

40 50 so

ESTIMATED OVULATION AGE IN DAYS

Fig. 1 Relation between estimated ovulation age and crown-rump length (675 embryos

from mothers with regular cycle).

weight in g). This curve, based on embryos whose ages were less than 40 days, is similar to that derived from Witschi’s data and strikingly discordant from ours.

c. Developmental horizon in relation to age. As shown in ﬁgure 3 a fairly wide range in developmental horizon was noted for any ovulation age. For instance, embryos at about 6 weeks of age had horizons ranging from 14-22. The heavy line in ﬁgure 3 indicates the mean curve for our specimens; the light line shows the mean age relative to horizons presented by Streeter (’42, "45, ’48, ’5l). Both are al— most parallel but our line represents about a 5-days-older age than Streeter’s for any given horizon and indicates that the

mean ovulation age for each horizon of our embryos was 5 days more than Streeter’s. Jirasek et al. (’66) reported a similar ﬁnding, but their -— Czechoslovakian—- embryos at horizons 17-23 were 8 days older than Streeter estimated his to be. Takakusu and Ban (’63) also pointed out that their —— Japanese --—--embryos at horizons 12-18 were on the average about 7-10 days older than Streeter’s. The stan~ dard described by Witschi (’56) with material mostly from the Carnegie collection was likewise several days older than Streeter’s. It might be pointed out that Streeter’s human series included pathological specimens obtained from spontaneous abortion or ectopic implantation and that

3000 0 I I-5 cases ’ 2500 O :6-IO cases 0 ‘:Il-20 cases o .: 22! cases E’ 2000 , ' E II O 0 9 g I500 Q . . . 5 . . ° . . O 0 0 0 0 0 0 0 m C C O C 0 now D Q C C O O 0 0 I C O C O C I 0 I O O 0 I O O O O I O C . . 50 C O C 0 I O O C 0 C O O 0 I I I 0 0 O 0 0 0 o o 0 0 0 0 0 0 0 0 O0 . 0 0 O O I I I 0 . . *1 go o o 0 0 I0 20 30 40 50 60

ESTIMATED OVULATION AGE IN DAYS Fig. 2 Relation between estimated ovulation age and body weight (675 embryos from

mothers with regular cycle).

these were supplemented with data obtained from rhesus monkeys.

(1. Developmental horizon in relation to crown-rump length (table 1). Again there was considerable Variation when horizon was plotted against length (ﬁg. 4), but this was not as great as with ovulation age. F or example, 10-mm embryos varied from horizons 15-18 and 20-mm ones from horizons 19-22, with coeﬂicients of variation of 4.31 and 3.58, respectively. Our curve indicates no remarkable differences from the observations of Streeter or Witschi in the earlier stages. Takakusu and Ban (’63) stated that the length of their specimens of horizons up to 18 almost coincided with that of Streeter’s. However, it is to be noted that in the speci mens of later horizons ours were consis— tently shorter than Streeter’s or Witschfs.

e. Developmental horizon in relation to body weight (table 1). Consistent with the ﬁndings shown in ﬁgure 4, ﬁgure 5 presents a certain ﬂuctuation with respect to the relation between developmental horizon and body weight of embryos, but again it is not as great as with age. The mean curve indicates some difference from the standard given by Witschi (’56). His ﬁgures of 20 mg for horizon 14 and 50 mg for horizon 15 fall nearly to our lowest level, while the ﬁgures of 1000 mg for horizon 19 and 1500 mg for horizons 2022 surpass our highest ones.

f. Crown-rump length and body weight (table 1). Figure 6 shows the relation ESTIMATED OVULATION AGE IN DAYS

embryos from mothers with regular cycle).

60

IO

NORMAL

TAB LE 1

AND

ABNORMAL HUMAN DEVELOPMENT

I5 I6

I7

285

Nlshlmura et al.(‘68)

Streefer ('42-'5!) ---—--— Witschi ('56)

I8 I9 20 2| 22 23

DEVELOPMENTAL H ORIZON

Fig. 3 Relation between developmental horizon and estimated ovulation age (675

Developmental horizon (Streeter) correlated with crown-rump length and body weight

Crown rump Body _ No. of length weight Horizon embryos (mm) (mg)

Mean i S.E. _ Mean i S.E. 12 14 3.9 i 0.21 7 i 0.2 13 38 5.1 i 0.15 16 i 0.2 14 163 6.8 i 0.08 47 3 1.3 15 142 8.0 i 0.07 80 i 1.9 16 145 9.2 i 0.08 125 i 2.8 17 48 11.5i0.13 211 i 6.5 18 24 13.5 i 0.36 339 i 13.8 19 29 15.9 i 0.30 521 i 16.5 20 35 19.2 i 0.26 790 i 25.5 21 18 21.1 i‘ 0.26 1020 i“ 33.3 22 13 22.8 i 0.53 1346 i 83.0 23 6 28.0 i 1.02 2098 i 100.7

between crown-rump length and body weight represented by logarithmic scale. The regression line in this ﬁgure has the following equation: Y = 2.517 X 10'‘ X17707 (X body weight in mg; Y= crown-rump length in mm), whereas Nakamura et al. (’55) gave a regression line with the following equation, based on their Japanese embryos ﬁxed with formalin: Y: 1.7249 X 10-1 X2-7941 (using same coordinates as above). Jirasek et al. (’66) presented the following equation for their fresh embryos: Y = 2.9026 X 10"1 X2-3523 (same coordinates). Both curves are similar to ours but that of Jirasek et al. (’66) is generally closer.

so '

Nishlmuro et cu. ('63)

Streeter ( '42- ‘till . . 25 ~ _ .

—--—-——-- Wlfschi ('56)

'8

CROWN'RUMP LENGTH IN MM 5 5

l2 l3 l4 I5 I6

I? I8

I9 20 2| 22 23

DEVELOPMENTAL HORIZON

Fig. 4 Relation between developmental horizon and crown-rump length (675 embryos

from mothers with regular cycle).

It was revealed that our standards are more or less different from those usually cited. Such differences cannot be attributed wholly to different sources of the specimens, nor can the inﬂuence of ﬁxation of the specimens or the unavoidable errors in measurement produce the striking differences that were observed in some instances. Remarkable variation was noted with respect to the relation between clinically estimated age and crown-rump length, body weight, or developmental horizon. This variability could to some extent be the result of inaccuracy of clinically rec ognized menstrual age. However, real differences in time of ovulation, interval between ovulation and implantation, or rate of early development surely must each play some role. Variations in the relation between developmental horizon and crownrump length or body weight, as well as between crown-rump length and body weight, were generally less than variability of the relation between age and horizon. Hence, it is concluded that crown-rump length and body Weight are generally more reliable indicators of developmental state than clinically estimated age.

one o :l- 5 cases ° 2500 O :6-IO cases 0 =ll-20 cases 0 . : ;2l cases (9 5 2000 ' E O ‘.3 S2 0 0 I500 i‘ 25 . . O on "' 0 I000 ' ' O O O O 0 0 O 500 _ . 0 0 Q 0 Q ' 0 0 ° --— 9 — I2 I25 I4 I5 I6 :7 us I9 20 21 22 23 DEVELOPMENTAL HORIZON

Fig. 5 Relation between developmental horizon and body weight (675 embryos from

mothers with regular cycle).

2. Abnormal development

All 1,213 embryos were examined for the occurrence of abnormal development.

a. Intrauterine embryonic death. The incidence of embryonic death, diagnosed macroscopically or by observation with stereomicroscope, in specimens obtained from women with or without a history of genital bleeding during pregnancy is shown in table 2. It is apparent that a signiﬁcantly greater percentage of dead embryos was retained in utero in women with than without genital bleeding. Ma-— ternal age and intrauterine embryonic death were unassociated (table 3).

b. External malformations. The incidence of malformed embryos at various developmental horizons is shown in table 4, and the incidence of each type of malformation in specimens that had developed sufﬁciently to allow detection of the respective defect, in table 5. Some of the gross malformations are described in table 6, where it is apparent that the variety of malformed embryos increased with advancing development. This is accounted for by new types of defects being. added as organogenesis proceeded. It is noteworthy that many abnormal embryos were found in women with a normal early preg288

§

aoov WEIGHT in MG 8

3 5 T

0 2 I--5 cases 0 : 6-IO cases Q : ll-20 cases

. : 221 cases

IO I5 20 25 30

CROWN‘-'RUMP LENGTH IN MM

Fig. 6 Relation between crown-rump length and body weight (675 embryos from mothers

with regular cycle).

nancy. In an unpublished study of our collection, the occurrence of external malformations Was determined in 5,667 partially damaged specimens, in which the following types were observed (ﬁgure in parentheses shows number of cases): cyclopia (2), cleft lip (5), myeloschisis (2), split hand (2), split foot (1), polydactyly (hand) (14), polydactyly (foot) (5), brachydactyly (hand) (1), oligodactyly (hand) (6), oligodactyly (foot) (1).

If the malformations in the undamaged and partially damaged specimens are considered together, it can be concluded that the embryos showed external defects primarily of the central nervous system, eyes, face, and extremities. It should be emphasized that because our specimens were very young it was not possible to detect malformations of external genitalia and

anal region or such defects as cleft palate and club foot, which are fairly common in infants. It is to be noted that no phocomelic specimen was found. This fact does not support Hellman’s (’66) hypothesis that thalidomide with its immunosuppressive action prevents the spontaneous abortion of fetuses with such a type of malformation.

A study of the concurrence of the abovementioned external malformations with visceral defects is being carried out by our group. Serial sections of whole specimens with external malformation revealed that among nine embryos six also had one or more internal anomalies of the digestive, urogenital, or endocrine systems. The details will be published elsewhere.

The incidence of several types of malformations in embryos were compared with corresponding data on infants (table 7) and it was found that the frequency of some defects in the embryos was far higher than in infants. Such facts were already pointed out by our group on the

TABLE 2 Incidence of death in utero Total No. and Group embryos % dead Without history of 1145 20 genital bleeding (1.75) With history of 68 9 genital bleeding (13.24) 1 Total 1213 29 (2.39)

1 P < 0.001 as compared with the above group.

289

basis of the examination of specimens collected to the end of 1965 (Nishimura et al., ’66). Such differences are also apparent when our data are compared with those on cleft lip, polydactyly, and spina biﬁda obtained in 24 centers in 16 countries by Stevenson et al. (’66).

The mean age (:2 S.E.) of the mothers of 13 malformed specimens, 30.00 i 1.58 (one specimen whose mother’s age was not recorded was omitted), was not signiﬁcantly diiferent from that of the 1,151 externally normal, live specimens, 29.98 2 0.18.

ACKNOWLEDGMENTS

We are indebted to 970 gynecologists in Kinki, Chubu, and Kanto area for their

TABLE 3 - ' ' l ‘n the s ecimens~ Age of mothers of live and dead embryos cooperatmn 111 Supp V1 g p = N f M t a] Dr. H. Yamamura, Dr. N. Matsuda, Mr. T. Category 03;; fag? Katsuya, B. S., and Mr. M. Terada, B. S., Mean -3 S.E. Live 1164 29.80 x 0.13 TABLE 5 Dead 26 31,15 : 1.17 Incidence of particular malformations, with respect to the horizon (H) at which each becomes recognizable TABLE 4 - ._ _ - 2 _n _._ Incidence of malformed embryos classiﬁed N0- and according to developmental stage ‘ eﬁggﬁs fr€%“f§§y . No_ of No. Frequency Malformation observed malformed Honzon embryos malformed (% ) Exencephaly 1135 (:1-1 12) 2 (0.17) 11-13 93 O — Myeloschisis 1163 (_‘-‘_:H I3) 2 (0.17) 14-16 765 3 0,39 Cyclopia 1095 (;H 14) 3 (0.26) 17-18 126 3 2,33 Cleft lip 204 (gH 19) 3 (1.47) 19-23 204 3 3,92 Split hand 819 (gH 15) 1 (0.12) Undetermined 25 —- —— Polydactyly (hand) 330 (gﬂ 17) 3 (0.90) Total 1213 14 1,15 Oligodactyly (hand) 330 (§H 17) 1 (0.30) TABLE 6 Anatomical description of the malformations Malformation Remarks Exencephaly Case 1 (H 1 20): Externally exposed brain tissue in parietal region Case 2 (H 20) : Similar ﬁnding in occipital region Myeloschisis Case 1 (H 14): Neural tube protruding in lumbosacral region Case 2 (H 16): Opened neural tube in sacral region appearing as a shallow depression Cyclopia Case 1 (H 19): Cebocephalic type accompanied by proboscis Case 2 (H 18): Single orbit with 2 eyeballs ' Case 3 (H 19): Single orbit with 2 eyeballs accompanied by proboscis Cleft lip Cases 1 and 2 (H 19): Unilateral cleft on the right side extending to naris Case 3 (H 21): Bilateral complete clefts accompanied by left split hand Split hand Case 1 (H 21): Deeply incised left hand plate with irregularly arranged digits, accompanied by cleft lip Polydactyly Case 1 (H 18): Small extra digit adjacent to first digit of left hand Case 2 (H 18): Similar ﬁnding to Case 1 on both hands Case 3 (H 23): Small extra digit as a branch of first digit of left hand Oligodactyly Case 1 (H 21 ): Missing first digit of left hand

1 H 2 Streeter horizon.

TABLE 7 Comparison of incidence of several external malformations between embryos and infants

n_

Incidence % (No. of cases and investigator)

Malformations

_

Exencephaly for embryos and anencephaly for infants

Myeloschisis

Cyclopia Cleft lip

Polydactyly (hand)

Embryos in our series

0.169 (1186) 0.172 1-2 (1163)

0.274 1 (1095) 1.471 1-2 (204)

0.909 1 (330)

Infants 0.062 (80435; Mitani, '54) 0.063 (63796; Neel, ’58)

0.019 (80435; Mitani, '54) 0.020 (63796; Neel, '58)

0.006 (80435; Mitani, ’54)

0.178 (80435; Mitani, ’54) 0.213 (63796; Neel, ’58)

0.098 (80435; Mitani, '54)

1 P < 0.01 as compared with Mitani’s data. 2 P < 0.01 as compared with Nee1’s data.

for their help in collecting the specimens; Miss C. Uwabe, B. S. and Miss Y. Kurotani, B. S., for their laboratory help; Mr. T. Kono, Miss M. Kono, and Miss M. Okada, for their recording. Also we wish to acknowledge the aid of Dr. J . R. Miller, Professor of Medical Genetics, Department of Pediatrics, University of British Columbia, Vancouver (Visiting Research Associate, Department of Anatomy, Kyoto University), and Dr. J. G. Wilson, Professor of Research Pediatrics and Anatomy, the Children’s Hospital Research Foundation, Cincinnati, in preparing the manuscript.

LITERATURE CITED

Aoki, H. 1962 On notiﬁed induced abortions in Japan. Annual Reports of the Institute of Population Problems, Ministry of Health and Welfare, Japan, 7: 45—52. (In Japanese.)

Arey, L. B. 1965 Developmental Anatomy. Saunders, Philadelphia.

Hellmann, K. 1966 Immunosuppression by thalidomide: Implications for teratology. Lancet, 1: 1136-1137.

Iffy, L., T. H. Shepard, A. Jakobovits, R. J. Lemire, and P. Kerner 1967 The rate of growth in young embryos of Streeter’s horizons XIII to XXIII. Acta Anat., 66: 178-186.

Jirasek, J. E., J. Uher, and M. Uhrova 1966 Water and nitrogen content of the body of young human embryos. Am. J. Obstet. Gynec., 96: 861-871.

Mitani, S. 1954 Malformations of the newborn infants. J. Jap. Obstet. Gynec. Soc., 1: 301-315. Nakamura, K., O. Saito, and T. Cho 1955 Measurements of early Japanese embryos. F.irst Report: Body length and body Weight.

Bull. Kobe Med. Coll., 6: 113-117. (In Japanese.) Neel, J. M. 1958 A study of major congenital

defects in Japanese infants. Am. J. Hum. Genet., 10: 398-445.

Nishimura, H., K. Takano, T. Tanimura, M. Yasuda, and T. Uchida 1966 High incidence of several malformations in the early human embryos as compared with infants. Biol. Neonat., 10: 93-107.

Patten, B. M. 1953 Human McGraw-Hill, New York.

Pearson, A. A., H. Nishimura, T. Tanimura, and R. W. Sauter 1968 Observations on the development of the external form of the Japanese embryo. Anat. Rec., 160: (Abstract) 489-490.

Section of Statistics in The Secretariat of Ministry of Health and Welfare, Japan. 1963 Statistics on Eugenie Protection in Japan, 1963. Ministry of Health and Welfare, Japan, Tokyo. (In Japanese.)

Shimamura, A. 1957 Age determination and physical measurements of Japanese embryo. Jap. J. Leg. Med., 11: 795-811. (In Japanese.)

Stevenson, A. C., H. A. Johnston, M. I. P. Stewart, and D. G. Golding 1966 Congenital malformations. Bull. Wld. Hlth. 0rg., 34, Suppl.: 1-127.

Streeter, G. L. 1942 Developmental horizons in human embryos. Description of age group XI, 13 to 20 somites, and age group XII, 21 to 29 somites. Carnegie Contr. Embryol., 30: 213-245.

1945 Developmental horizons in human

embryos. Description of age group XIII,

embryos about 4 or 5 millimeters long and age

group XIV, period of indentation of the lens

vesicle. Carnegie Contr. Embryol., 31: 29-63.

1948 Developmental horizons in human

embryos. Description of age groups XV, XVI,

XVII and XVIII, being the third issue of a sur vey of the Carnegie collection. Carnegie Contr.

Embryol., 32: 133-203.

1951 Developmental horizons in human embryos. Description of age groups XIX, XX, XXI, XXII and XXIII, being the ﬁfth issue of a survey of the Carnegie collection. Carnegie Contr. Ernbryol, 34: 165-196.

Takakusu, A., and T. Ban 1963 On the age of Japanese embryos. Med. J. Osaka Univ., 13.285-291.

Witschi, E. 1956 Development of Vertebrates. Saunders, Philadelphia.

Embryology.