588 CAROLINE B. THOMPSON 



tiniious ^vith the hypodermis, and its cells are modified hypo- 

 dermal cells. 



4. The inner cuticula is thinner at the margins and lacking 

 above the center of the lateral ocelH (fig. 13). The same is true 

 of the inner cuticula above the frontal gland (fig. 16). 



5. The cells of the frontal gland in the adult stages are secre- 

 tory in function and resemble other gland cells in structure, but 

 in the developing nymphal stages the cells produce no secretion 

 and bear a striking resemblance to the slender elongated visual 

 cells found in insect ocelli, Hesse ('01 b). According to our 

 theory this ontogenetic phase corresponds to a primitive, phylo- 

 genetic, condition. In this connection it is significant that in 

 very young n5rmphs the frontal gland is barely distinguishable 

 in surface views, while the brain and eyes are quite well devel- 

 oped. The farther study of these youngest nymphs by means 

 of sections may give additional evidence. 



6. The phylogeny of the termites upon the basis of the mor- 

 phology of the frontal gland suggested by Holmgren ('09) may 

 be referred to as additional evidence. Holmgren has divided 

 all adult sexual termites into three groups: The first group, 

 including all the higher termites, consists of those forms with a 

 sack-like frontal gland (sackformige Druse) and a fontanel open- 

 ing; the second group, including part of the lower termites, con- 

 sists of those forms with a plate-like nonglandular 'gland' (Fon- 

 tanellplatte) and no fontanel opening; the third group, includ- 

 ing the remainder of the lower termites, consists of forms in 

 which both the plate-Uke nonglandular 'gland' (Fontanellplatte) 

 and the fontanel opening are lacking. Holmgren beheves that 

 the sack-like fontanel gland and the nonglandular plate are 

 morphologically equivalent, and that both are derived from a 

 common ancestor without either gland or plate. He further 

 remarks that no fontanel is found in the Blattidae, which in 

 this respect are more nearly related to the lower termites. To 

 quote his exact words: 



Versuchen wir nun eine Phylogenie der Termiten auf dem Verhalten 

 der Fontanelle zu griinden, so m ichte erstens bermerkt werden, dass 

 diejenigen Formea, welche Fontanelldriisen (Fontanellplatte und sack- 



