THE DUCTLESS GLANDS 311 



which lined the original invagination that produced the thymus. This 

 view is incorrect, in so far as the original invagination was not hollow, 

 but solid, and therefore had no lining stratified epithelium. On the 

 other hand, this original invagination was composed of what would 

 otherwise have become stratified cells, and as some of these cells in the 

 central position give rise to the Hassall's cells, it might be argued that 

 these bodies represented a late and imperfect attempt on the ,art of the 

 gland to develop a lumen lined by a stratified epithelium. The process 

 might be compared with other invaginations which only develop their 

 lumen after their structure is well under way, as in the invaginated ner- 

 vous tube of some teleost fishes. 



Another remarkable development derived from the epithelial reticu- 

 lum of the thymus gland is seen in the form of small openings in single 

 cells, or formed by groups of cells, some of which develop cilia or cuticular 

 edges on such of their surfaces as bound these openings (Fig. 279, A). 

 The cilia are well formed and must undoubtedly be active during life. 

 The cavities in which they work are closed and often contain irregular 

 masses of some unknown secretion product. These ciliated openings 

 apparently have no function in which ciliary motion can bear a necessary 

 part, and they may be looked upon in the same light as Hassall's bodies, 

 as vestigial lumina which are lined with ciliated epithelium instead of 

 stratified. The few well-defined mucous cells found in the tissue must 

 also be regarded in the same light until some more definite function can 

 be proved to require them. 



The remaining and largest number of cells of this tissue are the ap- 

 parently amoeboid cells that resemble leucocytes of several varieties. 

 That some of these are derived from the mesoderm is undoubted. That 

 any of them are transformed epithelial elements is very improbable. 

 They have probably been acquired by the moving in of real lymph cells. 

 Their slight differences in structure and staining reactions are partly 

 responsible for the difference in texture between the cortex and medulla 

 of the lobules. 



We next shall examine the thyroid gland of a fish, Raja lavis, in 

 which the organ is typically developed. 



This organ consists of a series of lobules lying side by side in a mass 

 of vascular connective tissue. Each of these lobules is hollow and is 

 lined by a simple epithelium which is cuboidal in the smaller ones, and 

 high enough to be called columnar in others. The nuclei are round and 

 full and lie slightly proximad of the cell in the smaller lobules and well 

 toward the basement membrane in the larger ones (Fig. 280). The 

 cytoplasm is clear and shows several sorts of granules that probably 

 represent different stages in the elaboration of the secretion. The cells 

 have been divided into two classes on account of constant differences 



