208 Annals Entomological Society of America [Vol. IX, 



imagined that the protoplasm at the tip of the process with- 

 draws and the peripheral end of the tube unites with the walls 

 of the mouth of the reservoir. 



By the end of the third stage of development, the gland cells 

 are probably fully developed in size (compare Fig. 16, glc, and 

 Figs. 9 and 10, glc), but they certainly do not function as gland 

 cells until all the chitin is developed. The thicker the chitin 

 becomes, the longer the pore canals and conducting tubes 

 must be, and the farther the gland cells must migrate into the 

 lumen of the appendage. Up to the time of the fully developed 

 chitin, the gland cells, instead of producing a glandular secre- 

 tion, produce a chitinous secretion to form the conducting tubes. 

 Perhaps the withdrawal of the protoplasm from the tip of the 

 cell process keeps pace with the lengthening of the pore canal, so 

 that by the time the chitin is fully formed there is no protoplasm 

 surrounding the conducting tube. Now, instead of the gland 

 cells standing more or less vertically to the chitin, they lie 

 parallel to the chitin, and the conducting tubes no longer enter 

 the tips of the narrower ends of the cells, but they enter the 

 sides of the cells iisually about midway between the two ends. 



ib) Hypodermal glands without reservoirs. 



In a section through the articular membrane (Fig. 18, artm), 

 may be seen one or more small tubes. These are the conducting 

 tubes (C]) observed in chitinous preparations as shown in 

 Figure 8. They are so small and so transparent that they 

 were not noticed until after this piece of work was almost 

 completed. For this reason, it is not surprising to know that 

 other investigators have overlooked them. They have no res- 

 ervoirs. Each one pierces the chitin and terminates (pi) on the 

 surface of the flexible and comparatively thin articular 

 membrane. 



A layer of peculiar hypodermal cells had been noticed beneath 

 the articular membrane for some time before the conducting 

 tubes had been observed, but these cells were not given any 

 particular attention until the conducting tubes had been dis- 

 covered. A careful study of these cells (Fig. 18, glci) under the 

 highest lenses soon showed that they are gland cells like the 

 ones already described except they are several times smaller. 

 Since they are so small their internal structures can not be 

 studied critically. A conspicuous nucleus is always present, but 



