1914] Digestion and Digestive Epithelium in Insects al? 
the lighter cells the nidi (n) are to be found, mere rough groups 
of nuclei containing each a nucleolus and many granules or 
perhaps alveoles of varying sizes. The developing cells stain 
very slightly and frequently contain vacuoles. The active 
cells are longer, and, on account of the secretion which they 
contain, stain darkly. Both kinds have a fibrillar or palisade- 
like appearance basally, which extends as far as the nucleus. 
The inner portions of the active cells are alveolar, or possibly 
composed of a network, and contain many small highly re- 
fractive concretions. There are no distinct cell walls, but there 
is a periodical thickening of the fibrils, which give the cells a 
distinct appearance. The intima is moderately thick and 
traversed by pore canals. Frequently numerous droplets of 
secretion (sec) may be seen between this and the peritrophic 
membrane, and sometimes the secretion appears to be streaming 
from the active cells. In the region just behind the large crop 
the cells are smaller and more compact. 
The termites have a very peculiar digestive epithelium which 
perhaps can be correlated with their habit of feeding on dead 
wood. The stomach is bordered with from ten to twenty 
lobe-like projections, one of which is shown in Fig. 3. Each 
of these has at its base a nidus of many nuclei, and extending 
from this to the inner tip of the lobe, the cells overlap each 
other in a very curious scale-like manner which is evidently 
only a variation of the typical holocrine method of cell-forma- 
tion. It is noteworthy that this method should occur in the 
termites, which live in the wood they feed upon, and-at least 
have the opportunity of feeding continously, whether they 
actually do or not. 
In the order Hemiptera, I have only studied the rather 
abnormal Coccide (Fig. 4), which I have already mentioned. 
It looks here as though the large cells discharged their contents 
and were replaced by the smaller ones, which are formed by 
simple cell division, though I have not observed the process. 
In contrast to this arrangement we find in Myrmecophila, a 
small degenerate cricket (Orthoptera) inhabiting ants’ nests 
(Fig. 5), the typical nidi (n), with the regular wave-like ar- 
rangement of nuclei between them. Here, besides the intima 
with its pore canals, there are what appear to be cilia, and at 
their ends are small droplets of secretion. 
