MUSEUM OF COMPAliATlVH ZOOLOGY. 119 



tlio vsliuic by which the young "gars "are thickly covered. They arc 

 not, however, the only agents concerned in its production. 



The epidermal glands of the second kind are considerably smaller 

 tlian those already described ; they are nearly spherical, and have an 

 average diameter of from 18 to 25 fi. They do not appear to have been 

 recognized by previous observers. Some of these lie in the deeper 

 part of the epidermis, but much the greater number and the larger ones 

 occur near the surface, where many of them open. 



They stain deeply with Kleinenberg's and with Delafield's hsema- 

 toxylin, but in all other stains used, even in other hsematoxylin dyes, the 

 nuclei alone are colored, and in this condition they can be distinguished 

 from small glands of the first kind only by very careful observation. 

 This pi-obably accounts for their having been overlooked heretofore. In 

 Kleinenberg's ha^matoxylin the whole gland takes a blue tint, while the 

 wall pi"esents a reticulated appearance due to an irregular network of 

 lines of a much deeper blue color. This appearance is shown in Figure 

 2, a-e. The glands lying in the deeper part of the epidermis (d) take 

 less color than those situated just below the sui'face (c) ; those which 

 open at the surface vary much in the intensity of their coloring, the 

 differences doubtless being due to the varying amounts of mucin con- 

 tained in them. The distribution and peculiar staining qualities of 

 these glands show that they originate in the deeper part of the epidermis 

 and migrate to the surface, where their secretion is discharged. 



Hajmatoxylin has long been known to stain mucus deeply. Hoyer 

 ('90) found that basic stains are those which chiefly affect mucin, and 

 that hsematoxylin stains which contain alum act like basic stains. Of 

 the four ha3matoxylin dyes which I have used, Kleinenberg's (basic) and 

 Delafield's (alum) gave characteristic deep blue stains to these glands; 

 Boehmer's (alum) and Ehrlich's (acid), on the other hand, gave pure 

 nuclear stains.- Hoyer imputed certain failures of htematoxylin stains' 

 to act in their normal manner to lack of " ripeness," and it is possible 

 that this may be the reason for Boehmer's alum hsematoxylin not col- 

 oring the mucin in this case. 



iVt all events, I believe that the stains which I have obtained afford 

 sufficient ground for declaring that these cells act as glands, and secrete 

 the mucin element of the slime by which the young gar-pikes are thickly 

 covered. The other and larger glands first described, which take only a 

 nuclear stain with Kleinenberg's haematoxylin, must be considered to 

 have the function of secreting some component of slime other than 

 mucin. 



