40 Coelenterata. 



California. C. astuari has usually about 30, never more than 34, marginal, 

 and the same number of labial tentacles. The thin-walled column and tentacles 

 are highly contractile, the marginals may completely disappear during transport 

 to the laboratory, and they only gradually reappear in aquaria. These changes 

 are due to variations in the pressure of the fluid in the coelenteron and ten- 

 tacles. The siphonoglyph is broad, and on the opposite side of the stomodseum 

 there is a very narrow and faint depression. 6 mesenteries are long, 4 of them 

 reach the terminal (aboral) pore. The 10 longest mesenteries bear gonads, 

 while the alternating shorter ones bear much-coiled filaments along their edges. 

 The directives scarcely extend below the lower edge of the siphonoglyph. There 

 is no sign of the arrangement of the mesenteries in groups of four. True and 

 false acontia are wanting. Each gouad contains ova and sperms side by side 

 in separate follicles. - C. Benedeni has 90-100 marginal tentacles, arranged 

 in 3 cycles, and the same number of labial tentacles. There is a single siphono- 

 glyph, opposite which is a narrow groove. The mesenteries are of 4 orders of 

 different lengths. 18 (those of the first order) reach the vicinity of the terminal 

 pore; several of the 2nd order reach nearly as far. The directives extend a 

 short distance below the stomodseum but have neither mesenterial filaments nor 

 gonads. Filaments are found tightly coiled on the shorter mesenteries (the 

 3rd and 4th orders). Gonads are confined to the larger mesenteries (1st and 

 2nd orders) and are hermaphrodite. There are no true acontia, but there are 

 numerous, branching, filamentous processes of the mesenterial filaments similar 

 to those figured for C. membranaceus by Heider (1879). These occur on all 

 the mesenteries except the directives. On those of the 3rd and 4th orders they 

 are grouped at the oral end of each mesenterial filament, but on those of the 

 1st and 2nd orders they are scattered along almost the whole length of the 

 mesenteries beginning with a compact bunch situated just below the level of 

 the oral ends of the coiled filaments of the shorter mesenteries. On the mesen- 

 teries of the 3rd and 4th order there are also cni dor ages (van Beneden). 

 Each is a follicle packed with cnidoblasts which may be of several kinds. In 

 C. Benedeni they are very numerous on the bunched filaments, much less so on 

 the scattered filaments. The authors do not agree with van Beneden's view 

 that these aggregates (botrucnides) of cnidorages are morphologically acontia. - 

 In C. Johnsoni there are 105-108 marginal and 100 labial tentacles. The 

 siphonoglyph is as in G. B. The mesenteries are arranged in 4 orders; only 

 one pair reaches the vicinity of the terminal pore and the directives do not 

 extend below the lower level of the siphonoglyph. Hermaphrodite gonads are 

 borne on the mesenteries of the 1st and 2nd, and frequently also of the 3rd 

 orders. Filaments are found on the shorter mesenteries of the 3rd and 4th 

 orders. The acontia-like processes of the mesenterial filaments are similar in 

 structure to those of C. B. except that they have no cnidorages. This species 

 is most closely allied to the European membranaceus. See also Roule( 2 , 3 ). 

 According to Davis the sexually derived individual of Sagartia luciw is pro- 

 bably regularly hexamerous; of the 6 pairs of complete mesenteries, 2 are di- 

 rectives and are associated with a siphonoglyph. A secondary cycle of incom- 

 plete mesenteries, arranged in pairs, alternates with the pairs of the 1st cycle; 

 a 3rd cycle is usually present, and a 4th one may be present. Longitudinal 

 division is so common that such regular examples are rare. Fission, followed 

 by regeneration, plays an important part in the life history. Fission occurs in 

 endocoels, and in most cases, in complete endocoels. The fission plane tends te 

 lie at right angles to the major transverse axis, producing bilaterally symmet- 

 rical pieces but with little regard to accurate halving of the animal. In re- 



