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



epithelium must evidently correspond with that of the rete 

 mirabile. It need hardly be pointed out that the gross ana- 

 tomical simplicity or complexity of the "red body" by no means 

 always corresponds with the histological simplicity or complexity 

 of the gland epithelium, and that this latter affords the only true 

 basis for a classification of " red bodies." One of the most simple 

 (though not the most primitive) conditions of the rete mirabile 

 is when it is single and compact*. In this condition it may be 

 separate from the glandular epithelium and ovoid in shape 

 (Ophichthys iviberbis, fig. 9, e.g.), or contiguous with the glandular 

 epithelium and circular in form (Gobius, fig. 22, e. g.), conical 

 [Cepola, fig. 67, e. g.) or fan-shaped (Ophidium, fig. 48, e. g.), &c. 

 In some cases the rete mirabile is divided into two, and here 

 these are separate from the glandular epithelium and ovoid in 

 form (Angitilla, text-fig. 53, j\Ji/ruSj Mztrcsna, Ophichthys serpens, 

 (fee). In other cases the rete mirabile is divided up into several 

 distinct par-ts, which often have a radiate arrangement : e. g. in 

 Peristethas the rete is split up into some ten radiating strands 

 (fig. 37 and text-fig. 57) and in some species of Gobius into seven 

 or more. In most cases, however, the initial artery and vein 

 supplying the bladder give off numerous arteries and veins in pairs 

 (an artery and a vein to each pair, of course) at intervals so as to 

 give rise either (a) to continuous sheets of rete mirabile, so to 

 speak, very often two in number, one on each side, as in Corvina 

 nigra (fig. 59), Sargus rondeletii (fig. 60), and Smaris vulgaris 

 (fig. 45), or (6) to sheets divided up in variable degrees, as e.g. in 

 Box boops, where the sheets are divided into outer and inner on 

 each side (figs. 51, 52), or in Zeus faber, where there are three 

 separate sheets — anterior, median, and posterior — on each side 

 (fig. 62), or (c) to numerous separate small strands or tufts of rete 

 mirabile, each strand supplying a small area of glandular 

 epithelium : thus, in Co7'is jidis (figs. 56, 57) the rete is seen 

 under the microscope to be subdivided into many small strands 

 supplying the small area occupied by the gas gland ; in Trigla 

 hirundo also a similar condition exists, but here, indeed, the twigs 

 extend nearly over the whole interior of the bladder cavity (fig. 43) 

 instead of being restricted to a small area ; in other cases where 

 the tufts are larger and more separated, these can be seen with 

 the naked eye, as, e. g., in Atherina hepsetus (fig. 53) and Perca 

 (fio'. 69) and in the well-known case of the Cod (fig. 66), in which 

 the subdivision of the rete mirabile reaches its maximum. 



Whether the condition of the rete mii^abile with which we 

 started, i. e. as a single large well-defined body, separate from the 



* Reis (64) supposes that the horseshoe-shaped gland of Ophidium, e. g. (see my 

 fig. 48 and compare it with figs. 59, 60, and 62), is the form of gas gland and rete 

 from which all others are derivable, but it seems to me that the simple oval rete of 

 Syngnathus in connection with an extensive area of glandiilar epithelium, is still 

 more simple ; the most primitive condition of the epithelium and associated blood 

 vessels is, of course, that of many freshwater teleosts in which practicallj' the whole 

 of the internal unilaminar lining-epithelium of the bladder constitutes the gas gland, 

 and the arterial and venous capillaries have not become associated to form a rete. 

 In general it may be said that the more ditl'usethe rete and gland the more primitive 

 their nature. 



