454 BULLETIN 82, UNITED STATES NATIONAL, MUSEUM. 



It is probable (though not proved) that only the true sense cells carry cilia. 



Attachment disk and vestibular invagination. — The attachment disk, first 

 demonstrable at a considerably younger stage, now appears elongated and deepened. 

 Within its compass in the fully formed embryo the prismatic cells show a fine 

 ciliation, but the cilia are markedly shorter than those of the apical pit. Later the 

 cells diAnde and become wandlike, but they never become so fine as those of the 

 apical pit. No subepithelial ganglion cells are formed here, and apparently there 

 are no sense cells. Such ganglion cells as occur are confined to the border zone 

 and appear to have arisen through the division of the ganglion cells originating 

 from the ectoderm of the apical pit. 



The nuclei of the cells in the attachment pit are mostly somewhat larger 

 and more elongated than those in the sense cells, and in liEematoxylin stain a little 

 less deeply. They lie at different heights in the inner half of the cell, only 

 occasionally in the outer half. 



Sometimes the cells at the base of the pit are very long, almost a third again 

 as long as the sense cells in the apical pit. 



Just before and during the fixation of the larva the glandular character of 

 this region becomes more and more marked; the cilia disappear, the cells become 

 mostly somewhat shorter and thicker and give off exteriorly a layer of homoge- 

 neous secretion, which accomplishes the fixation. 



In the larva just liberated the vestibular invagination extends over the greater 

 part of the ventral side. As in the younger stages, it is deepest anteriorly, becom- 

 ing shallower posteriorly where it is bounded by the anterior border of the third 

 ciliated band. Very soon the ventral median portion of the third band becomes 

 involved in the extending invagination, which reaches to the fourth bow. 



The cells of the vestibular floor are long and almost wand-shaped, and their 

 nuclei lie at different heights, though in the anterior median portion almost 

 exclusively in the inner end near the body cavity, from which, as from the mesen- 

 chyme cells, these cells are sharply cut off. 



On their outer ends tlie cells carr^y cilia which are embedded in a cuticlelike 

 layer, so that the section shows a certain similarity to a section through the ciliated 

 rings. 



The whole invaginated region is lined with a single layer of cells. Such cells 

 as do not reach across the entire thickness have probably just arisen through 

 division and have not as yet reached their ultimate length. 



Because of the enlargement of the surface which this portion of the ectoderm 

 wall undergoes, active cell division must occur. In the posterior part of the begin- 

 nings of the vestibule, where at first only a slight flattening is found, the ectoderm 

 is noticeably thickened and shows numerous dividing cells. Later the median 

 portion sinks in, and this posterior part assumes the character of the more anterior 

 sections. 



If the larva remains free swimming for more than 24 hours important changes 

 occur in the vestibular invagination. If it attaches itself immediately these 

 processes take place in the fixed stage after the disappearance of the ciliated bands. 



