Respiration in Invertebrate Animals. 299 



markable regularity in size and outline. They are not detectable 

 on the internal walls of the parenchymatous passages which 

 coincide with the capillary segment. The blood-channels are 

 therefore here imparietal. The epithelium ceases where the spe- 

 cial boundary of the artery ceases ; it begins again at the limit 

 denoting the origin of the true veins. This hexagonally-celled 

 epithelium is the prevailing envelope or lining of all organs 

 and cavities in the Ci-ustacea. It betrays no other diversities 

 than those which depend upon the size and distinctness of the 

 contained cell-granules. The cells are 7iever furnished with a 

 nucleus. The granules are in the interior of the cells, and adhere 

 internally to the cell-wall. A different opinion is expressed how- 

 ever by Professor Quekett : he describes the granules as belong- 

 ing to the underlying structures. The error of this description 

 may be placed beyond doubt by the reagency of acetic acid. 

 Dr. Carpenter denies the existence of cells in the epidermis. In 

 the adult Crab, for some time after the moulting has taken 

 place, that is, after the shell has become hard, it is, as this 

 author states, impossible to detect the cellular arrangement of 

 the membrane exterior to the calcareous layer. The cells seem 

 to have been mechanically worn away. Soon after the moulting 

 however, the presence of cells in the epidermis of the carapace, 

 for instance, admits of easy demonstration. 



The hexagonally-celled epithelium is an element of varied use 

 and great importance in the crustacean organism ; it constitutes 

 real boundaries everywhere of the extreme or capillary circula- 

 tion (PI. XVIII. fig. 5). Plates {a, h) are formed by the ap- 

 position of its constituent cells laid accurately edge to edge : 

 these plates are united by interposed islets or patches of pai-en- 

 chyma (PI. XVIII. fig. 1, a). Between the latter are left 

 large, irregularly and angularly bounded passages, traversed by 

 the extreme blood-currents (fig. 1, b). The islets of paren- 

 chyma consist of a variable number of nucleated cells, filled 

 obviously with the fluid elements of the blood. The groups 

 differ in size and outline in different organs. They are some- 

 times embraced by a common capsule : in such a case the latter 

 would constitute the real boundary of the blood-channel. As 

 such a capsule is the independent envelope of a detached group 

 of cells, rather than the continuous boundary of a conduit, how- 

 ever irregular in form and outline, it cannot with any anatomical 

 propriety be defined as the wall of the latter. The epidermal 

 plates, between which the blood- passages are disposed, are in- 

 flexible, firm, non-contractile. " Membranous parietes'^ of blood- 

 vessels adherent internally to these plates could not contract 

 upon the contained fluid without approximating the plates. 

 Such an effect would imply a strong muscular effort. IVo 



