MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 521 



appeared singly, or only two or thi-ee at a time. Probably the larvte are able to made their exit 

 through the tips of the tentacles, as well as through the mouth, though owing to the rapiditj^ 

 with which the process takes place J have never been able to assure myself of this method, even 

 when, as in the case of Siderastrscd, colonies have been watched for hours. Von Koch states 

 (ISyT) that the larvffi of Caryophyll'ia cyathus pass through the tips of the tentacles. The various 

 polyps of a colony contiiuie to give out larva? for several days, or even for a week or two, and 

 then the supply ceases, or for some time one or two individuals may appear at long intervals. 



Development appears to proceed equalh' within most of the polyps in any colony. In the 

 fertile colonies of J^avia or Siderustrxa the majoritj' of the polyps contained larvaj all at the 

 same stage, and in I'orite.s such was the case with most of the polyps within any restricted area. 



On first extrusion the larva- of corals are spheroidal, oval, pear-shaped, or elongated rod-like 

 bodies, varying from 1 to 3 nun. in length, the outer surface uniformly ciliated throughout. The 

 vai'ious forms assumed by the different larvfe of Favia fragum. extruded about the same time, 

 indicate the more usual shapes (PI. XllI, tigs. 96-100). Lacaze-Duthiers (1873) figures the larva 

 of Axiroides cd/i/eular/s as elongated and assuming a spiral form, von Koch (1897) that of Caryo- 

 phyllia cyathus as pear-shaped. The individual larva, however, often manifests the iwwer of 

 retraction and of altering its shape, so as to be at different times oval, pear-shaped, spheroidal, 

 or riattened and cake-like. In the majority of cases, one end, usually the anterior during progres- 

 sion, is much broader than the other, though when the larva has been extruded for some time 

 these relations are frequently reversed. Thus, the pear-shaped larva of Favia fragum has at 

 first the broad end at the aboral forward pole, but later the broad end is oral or posterior 

 and the narrow end is aboral. A similar alteration of form has been noticed in Siderastrma 

 nidiaiix, but most of the larva? of this species have a narrow aboral and a swollen oral pole from 

 the beginning. Twin larviv, with two oral extremities and one aboral, have been extruded by 

 polyps of the species just mentioned. No coral larva has 3'et been described in which the aboral 

 extremity bears the tuft of larger, less mobile cilia sometimes met with in Actinians. 



Coral larvfe are able to swim about either immediately on extrusion or shorth' after. For the 

 iirst few minutes they may remain motionless, either at the surface of the water or on the bottom 

 of the vessel, then, cilia having been formed, active movements of both rotation and translation 

 commence. Some gyrate throughout the depth of the water, coming to rest from time to time; 

 others remain nearer the surface or accumulate around the sides of the vessel. Within one to 

 three days a few of the more vigorous examples would become attached to the sides of the 

 glass vessel or other object, and remain thus for some time, then become active again, and 

 afterwards refix themselves. But the great majorit}' seemed unable ever to settle, and continued 

 alive for daj's or weeks, without much motion, and apparently without undergoing further 

 development. If fixation did not take place during the first two or three days it was never found 

 to occur after, though some of Wilson's larva? of Maiiicijia settled after swimming around for 

 three weeks. 



When first extruded, coral larvw are dense and opaque, and either colorless or slightly 

 brown; afterwards they may become slightly distended, and as a result the wall appears thinner and 

 inoi-e nearly transparent. Occasionally the iarvie are set free in a distended form, when they are 

 more transparent from the beginning. The alteration from the opaque to the more transparent 

 condition is brought about by an important change in the internal endodermal tissue, described 

 below. Under the microscope the colorless or nearh' colorless outer ectoderm can be distin- 

 guished from the inner endoderm, and in all species examined the latter bears zooxanthellse, 

 which give a yellowish color to the internal mass. 



The oral pole, whether narrow or t)road, is usually darkly colored externallj*. Examination 

 of the living larvi^? under the microscope, and also by means of sections afterwards, reveals that 

 the color is due to the presence of numerous zooxanthella' or yellow cells toward the oral end. 

 Usually these occur within the ectoderm cells, but sometimes, as in hnpliyUia dipsacea (PI. XXV, 

 fig. 165), they crowd the endoderm cells around the oral e.xtremity, and the ectoderm contains 

 I'oniparatively few. In both cases the zooxanthellte are densely aggregated toward the apex, 

 whic'h in consequence is the darkest area. In general the alga^ are sparingly distributed 



