208 DR. W. N. F. WOODLAND ON THE 



intermediate in character between the folded and massive types, 

 but we may still retain these two categories, since we have reason 

 to sujjpose that all of these three supposed transitional forms of gas 

 gland really belong to the folded type and that they have only 

 secondarily acquired features characteristic of the massive type 

 by the splitting-ofF of a superficial cell-layer covering the surface 

 of the gland, by close apposition and anastomosis of the elongated 

 folds, and by occasional local prolifei-ation of the cells ; in other 

 words, until their development is known, we may provisionally 

 distinguish between truly or primarily massive glands which 

 never exhibit the folded arrangement of cells at any stage of 

 development, and ps-eudo- or secondarily-massive glands which are 

 secondarily derived from the folded type*. To me Reis's sugges- 

 tion that all massive glands have been derived from folded glands 

 seems very improbable, since on this view the folded condition 

 should occur as a stage in their development, and we have seen 

 that this is not the case in at least one typical genus {Atherina). 

 It seems more probable that both folded and massive types have 

 originated independently from the simple unilaminar condition 

 which must in all cases form the starting-point of development 

 both in ontogeny and phylogeny. 



CoRis JULis t (vulgaris Flem.), 



The " red body " of Coris julis has the circular shape seen in 

 figure 56 (PI. VI.), and the rete mirabile supplying the gland is 

 noteworthy for the way in which it is broken up into small divisions 

 (PL VII. fig. 57). The glandular epithelium (fig. 58), which is 



* Reis's classification of gas gland epithelia (64) seems to me confinnatovy of 

 this view. Her classification is as follows : (1) entireij' nnilaminar epithelium 

 folded into simple tubular outpushings ; (2) the epithelial folds are umcli folded 

 and branched [Keis has made the mistake of including in this group the gas gland 

 epithelium of Coi-vina ; in mj' preparations of Corvina niyra the epithelium is of the 

 most_ typical massive type] ; (3) the epithelium is in parts folded and in parts 

 multilaminar [I cannot agree to the inclusion of Sijngnathus in this category. I 

 have not examined Reis's other example — Hipponampits] ; (4) the compact or massive 

 glands of Sargtis, Charax, and other genera. Reis regards these four types as a 

 developmental series — a mistake I have corrected in the text. Reis also remarks 

 that only in the first three types are the ducts obviously the spaces between the 

 epithelial folds or tubular outgrowths ; in the fourth there exist no such intertubular 

 ducts, intercellular crack-like channels only being present. This last statement 

 seems to me to be confirmatory of my contention. 



Reis makes another statement which, as Oppel (58) points out, contravenes all 

 probability. This statement is that whereas in the folded type of gland the 

 individual gland cell only presents one surface towards the duct lumen (the others 

 being contiguous with other cells and with a blood-vessel), in the massive gland the 

 individual gland cell presents all its surfaces towards the intercellular lumina pene- 

 trating the mass of the gland, i. e. the intercellular lumina, according to Reis, 

 surround each cell. Reis suggests that this supposed feature is for the purpose of 

 facilitating the exit of the gas through the gland into the bladder. As Oppel 

 justly I'emarks, how can a cell thus shut off from all vascular tissue obtain nourish- 

 ment ? So far as my observations go, there is no reason to suppose that a single cell 

 of the massive gland is not in contact with a capillary. 1 cannot, however, agree 

 with Oppel's groundless suggestion that tubular and massive glands possess difierent 

 functions, tubular glands being secretory and massive absorptive. Jf this be so, from 

 whence do fish with massive glands obtain their bladder oxygen ? 



t Mr. Tate Regan informs me that, according to Holt, " Coris giofrcdi " is merely 

 the female form of C.Jtilis. 



