BY R. J. TILLYARD. 433 



We have already alluded to the fact that these capillaries form 

 complete loops. This is most important, since it points strongly 

 in favour of the view that gas must enter equally readily at all 

 points of the circumference of the capillary. We should note, 

 further, that each capillary passes up one wall of the gill and 

 down the other. Thus it is arranged to the greatest advantage 

 for the reception of oxygen from the water, which bathes both 

 sides of the gill. Finally, as there is no spiral thread in these 

 capillaries, and their walls are excessively fine, they oflfer no 

 obstacle to the free passage of gas into their cavities by diffusion. 



It is well known that chitin, which is the substance forming 

 the cuticle of the gill and the intima (endotrachea) of the capil- 

 lary tube, is a colloid substance which admits of the passage 

 through it of gases by diffusion, and is particularly partial to 

 carbonic acid gas. It is generally supposed (and the experiments 

 of Dewitz(2) support this view) that chitin possesses some special 

 power of absorbing carbonic acid gas and giving it out again. 

 Hence we might expect that a thick layer of chitin would facili- 

 tate the passage of this gas, whereas, for other gases, such as 

 oxygen or nitrogen, a thin layer would be necessary. This is 

 significant when we contrast the thickness of the body-cuticle, 

 through which carbonic acid gas is now generally believed to 

 pass outwards, with the delicate layer covering the gills, through 

 which we claim that oxygen passes in by diffusion. 



It may be urged that the arguments applied above, in favour 

 of the Diffusion-Theory, would be just as applicable to a theory 

 which would substitute some chemical means of extraction in 

 place of diffusion. May not the gill-epithelium or its pigment 

 be an active oxygenating agent 1 To this the reply is that, if 

 the gill-epithelium possessed this power, (i.) it would be sufficient 

 for the capillaries to lie in the hsemocoele beneath the epithelium, 

 since the oxygen would be given out" to them by the inner border 

 of the epithelium; the fact, that they are pushed in towards the 

 cuticle, speaks in favour of diffusion; (ii.) the thinning down of 

 the epithelium, and the fusion of its cells into a single proto- 

 plasmic mass or syncytium, would not be an advantage in this 

 case, since a thicker layer of separate, chemically-active cells 



2 



