MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 523 



one another, and produced angulated outlines. In one instance a colony thus produced consisted 

 of thirty-two primaril}' free larva?, in another twelve larvte associated themselves, in a third 

 seven, while several groups of three or four poh'j)s were formed. Some of these colonies lived in 

 small aquaria for three or four months, during which time the tentacles and skeleton appeared. 

 To all appearances they would, under natural conditions, have given rise to actual colonies, 

 indistinguishable from ordinaiy colonies produced by gemmation. (See foot-note, p. 495.) 



Of the numerous larvre extruded l)y corals comparatively few seem to settle, and in aquaria 

 the greatest difBculty and uncertainty are experienced in securing permanently fixed individuals. 

 Sometimes a number will l)ecome fixed, while under what appear to be exactly similar circum- 

 stances fixation seems impossible. In addition to the somewhat unnatural conditions under which 

 the larva may be placed in aquaria, it seems not improbable that the ripeness of the larva? for 

 settling may also be a factor. In my experience, if fixation be not efl'ected within two or 

 three days after extrusion, it does not take place afterwards. The larvas will then remain resting 

 or slowly swimming about for an indefinite time, apparently undergoing no development what- 

 ever. Larvie of .V. rach'anx have been kept thus for a period of twenty daj's. 



Difterent measures were employed to provide the larvse with suitable surfaces for fixation, 

 and at the same time permit of their examination later. Glass dishes and small pebbles were 

 placed in the vessels, and cover glasses fioated vertically V)y means of pieces of cork. These 

 jDrovisions, however, were of little service. The most favorable position appeared to be the sides 

 and bottom of the glass vessels in which the colonies were living. The larvfe being properly 

 settled, the vessel was broken with care, and the fragments bearing the larva? distributed to other 

 vessels in which coral colonies were already established. Many larva? were .secured in this way, 

 fixed to transparent pieces of glass, and could be taken out at any time, and examined in small 

 glass dishes as ti'ansparent objects under the microscope. 



Once the larvie were fixed, they appeared quite vigorous and hard}-, and continued their 

 growth even under unfavorable conditions; while larvae which remained unfixed, though kept 

 alive and active for several weeks, never increased in size or underwent development in any waj^. 



Ectoderm. — The larval ectoderm is very broad compared with the same layer in the adult 

 polyp. In section the ectoderm of the larva of Agaricia measures 0.1 mm., and that of Favia 

 fragion 0.08 mm. Most of the usual Anthozoan cellular elements are alreadj' differentiated at or 

 before extrusion; gland cells, nematoblasts, supporting cells, and nervous elements occur, but 

 no muscular fibrils have been recognized. No observations have been made on fresh macerated 

 material, but the various cells separated somewhat freely from certain of the specimens preserved 

 in formalin. 



Both transverse and vortical sections of the ectoderm exhibit certain zones characterized bv 

 diflerentiations in the cellular constituents (fig. 165). The greater number of the nuclei are 

 aggregated about the middle of the layer, and inwardly they occur in diminished numbers as far 

 as the mesogla?a. B\' reason of the deeply-staining character of the nuclei their zone of 

 distribution stands out strongly in moderately thick sections, and macerations show that it 

 comprises the nuclei belonging to the supporting cells; the more deeply situated nuclei are 

 those of the gland cells, developing nematocysts, and nervous elements. 



The outer half of the ectoderm comprises the swollen portion of the gland cells and the mature 

 nematocysts. embedded, as it were, in a matrix of supporting cells; the margin frequently 

 shows the swollen bases of the cilia, which stain ver}' strongly in method blue. 



The inner zone is not well defined, and in early larvfe is usuallj' characterized by the presence 

 of large numbers of developing nematoblasts. These stain deeply, appear nearly homogeneous, 

 but with a nucleus to one side, and are arranged irregularly at all angles to the other constituents; 

 as thej' mature they migrate peripherally, and become arranged at right angles to the surface of 

 the layer. 



The gland cells ai'e a ver^' important constituent of the larval ectoderm, and their contents 

 are nearly always finely granular, and usually remain unstained. Sometimes the vacuolar part 

 extends nearly, if not altogether, across the layer, and on preservation the larvi« often throw out 

 large quantities of mucus, when the cells become clear. 



