Dissection ofa specimen of Latimeria chalumnae in the American Museum of Natural History revealed that it is a gravid female containing five advanced young, averaging 317.8 millimeters long. Each has a large yolk sac with no apparent connection to the surrounding oviducal wall. We conclude that Latimeria is ovoviviparous. Fig. 2. Yolk-sac young of Latimneria chalurnnae, the sccond of five found in a temale 1.6 m long at the American Museum of Natural HistorN. Unretouched photograph shows fins and tail as compressed hs oviducal walls. Total length, 322 mm. The specimen ot Latimeria in the American Museum of Natural History measures 1.6 m in total length; its weight at the time ot capture was reported to be 65 kg (9). The anatomy of the head of this specimen has been investigated (10), but its viscera were not dissected until recently, when samples o' hemopoictic tissues were taken. During the course of the dissection, tive advanced young were discovered Iving free in the right oviduct. Millot and Anthony (1, 2) have observed that only the right oviduct is f'unctional in Latimeria. As indicated in Fig. 1, all of the young were situated with thetr heads directed away from the urogenital orifice. The following observations were made on four of' the young that were removed from the oviduct, the fifth having been lef't in situ. The young resemble miniature adults, differing most noticeably in the possession of' a volk sac, the relatively larger eves, and a more declivous profile (Fig. 2). Total length ranges from 301 to 327 mm, with an average of3 17.8 mm. The maximum diameter of the yolk sac ranges from 80 to 129 mm; the largest fish has the smallest yolk sac, and vice versa. The volksac stalk is broad, extending from the base of the pectoral fins approximately twothirds of the distance to the base of the pelvic tins. Scales and fins appear to be fully developed in all four young, but they lack the denticles (odontodes) of the adult. The gravid f'emale under discussion was caught in January, the same month in which females with ovulated eggs have been taken. This suggests that gestation may require more than a year not an unexpected length of time in view of the tremendous size of the ripe egg. Millot and Anthony (1) showed that the male of' Latimeria possesses a cloaca which contains a urogenital papilla and is flanked externally by two pairs of erectile caruncles. Since internal fertilization of the 1106 female must occur, it seems likely that the cloaca functions as an eversible copulatory organ, in a manner reminiscent of'the situation in some birds and gymnophiones (11). The caruncles perhaps serve as claspers. Similar suggestions have been made by Griffith and Thomson (6, 12). Paleoichthvologists usually consider the coelacanths to be most closely related to the Paleozoic rhipidistians, and the latter in turn to be the fishes most closely related to the tetrapods (13). Nothing is known about reproduction in the rhipidistians, and there is no way of' knowing whether the ovoviviparous condition in Latimeria is unique or whether it is shared by other crossopterygians and primitive tetrapods. Among other major groups of bony fishes, living lungfishes are oviparous, as are all living actinopterygians except II families (less than 5 percent) that exhibit viviparitv or ovoviviparity as a derived condition or specialization. Although the ovoviviparity of Latimeria sheds no light on the reproductive mode of primitive osteichthyans, including crossopterygians, it does indicate that all the information we now have about reproduction in the fossil coelacanths is consistent with the hypothesis that they were ovoviviparous. C. LAVI iT SMITH American Museum of Natural History, New York 10024 CHARI FS S. RAND Department of Biolog, Long Island Universit v, Zeckendorf C ampus, Brook/in, New York 11201 BOBB SCHAFU II R, JANIE S W. ATZ A merican Museunm oJ Natural History References and Notes, J. Millot and.l. Anthonsr, Bull. Mu.m. Natl. Hist. Nat. Ser. 2 32, 287 (1960). 2. _, C. R. Hekd. Seances Acad. Sci. 251. 442 (1960). 3. J. Anthony and J. Millot, C. R. HeAd. Seances Acad. Sci. Ser. D Sci. NVat. 274, 1925 (1972). 4. J. Millot and J. Anthornv, ibid. 276, 2447 (1973). S. Sei. Nature Env!iron. No 121 (1974). p 3. 6. R. W. Gritlith and K. S. Thomson, Nature (Lond.I 242, 617 (1973). 7. D. M. S. 'satson, Proc. Zool. Soc. Land. 1927, 4_3 (1927). 8. H. P. Schultze, Nat. New Biol. 236,90 (1972). 9. The specimen, which is specimen 26 in the summars table of coelacanth captures prepared bs J. Millot, J. Anthony, aind D. Robineau [Bull. ius. NVail. Hist. Nat. Ser. 3 (No. 53), 533 (1972)], was caught off Mutsamudu, Anjouan Island, in 1962. G. W. Garrouste, ai physician then lising in Anjouan, arrangcd for its acquisition bs the American Museum of Natural Histors. In the same table specimen 65 is incorrrcctly recorded as having zalso been sent to the American Museum. 10. For example, bs G. J. Nelson. Bull. Ani. Mlu. Nat. Hist. 141, 475 (1969), Copeia 1970, 468 (1970), Zool. J. Linn. Soa. 53 (Suppl. I), 333 (1973). 11. S. B. McDowell, personal communication. 12. In the light of the present discovery, (iritlith and Thomson also may have been correct in suggesting that the isolated volk-sac larvae described by Schultze were prematurels released from . stressed female. Such behastior is often seen in ovoviviparous sharks, rays, and teleosts. 13. J. A. Mos-Thomas .ind R. S. Miles, Palaeoaoic Fishes (Saunders, Philadelphia, cd. 2, 1971). 14. We thank C. G. Schlcifer ('or the drawing, C. Tarka for the photograph, and Dr. G. J. Nelson for his cogent comments on the manuscript.