As the glabella of most trilobites is regularly convex, there must lie beneath it an organ running from front to back, which presses the bases of the cephalic legs away from each other and down from the dorsal test. An organ so extensive and unpaired, running thus from front to back, can, among the Arthropoda, be regarded only as an alimentary canal, for the swellings of the cephalic ganglia and the heart are by far too small to produce such striking elevations on the front and upper surface of the glabella. The canal might then have consisted of a gizzard belonging to the oesophagus, and astomach proper or main digestive canal. ... Among the trilobites there are two pairs of vessels on both sides of the glabella which have precisely the same position with reference to the supposed course of the alimentary canal as the ducts of the hepatic lobes in _Limulus_. One observes in numerous trilobites, although in different degrees of clearness and under various modifications, a dendritic marking of the inner surface of the cheeks which takes its rise at the lateral margins of the glabella and spreads thence like a bush over the entire surface of the cheeks. Exactly the same position is taken by the richly branched hepatic lobes of _Limulus_ on the lower surface of the head shield; a fact of special weight in favor of the homology and similar significance of the two phenomena, is that in the trilobites also, the anterior of the two main ducts is the larger, the posterior the smaller. The striking similarity of the two structures is shown by a comparison of the head shield of _Eurycare_ [_Elyx_] from the Cambrian of Sweden, in which the course of the canals is shown with remarkable clearness [with those of _Limulus_]. I have been able to convince myself that the existence of the two canals on each side is also the rule in other genera, even though the posterior pair is frequently but feebly developed or completely obscured by the anterior pair. In _Dionide formosa_, for example, I find only the anterior pair, which is very large and divided into two principal branches. From all these considerations it seems to me no longer doubtful that the median elevation was caused by the stomach and gizzard, and that the cheeks have principally served to cover the hepatic appendages of the alimentary canal.