90 DEVELOPMENT OF VESSELS. 



the islets of parenchyma, and to enter the blood-current, where 

 tliey are converted into blood-corpuscles. Such phenomena as 

 these, however, are only apparently contradictory to the well- 

 kncvvn structure of the apparatus when complete. The very next 

 step clears up this point. This step is the formation of an endo- 

 thelial lining to the vessels, which is justly regarded by Eherth 

 (StricJcer, ^'Manual of Histology,*' Syd. 'Soc. transl.," i. 264) 

 as the essential basis, the only universal and characteristic ele- 

 ment of the vascular wall. With the formation of the endo- 

 thelial tube, w^hich, as the reader is aware, consists of a very 

 delicate membrane, made up of elongated, polygonal, nucleated 

 cells, all detachment of cells from the inner surface of the vascular 

 Avail is at an end, and a definite boundary-line is established, 

 beyond which the parenchyma, in the broad sense of the word, 

 may be said to begin, wdiether it consist, as in the present case, 

 of embryonic tissue, or of connective substances, or of connective 

 tissue with muscular fibres, &c. 



The completion of the endothelial tube brings the process of 

 primary vascularisation to a close. Should a new vascular loop 

 be required, it can only be formed by a protrusion of the endo- 

 thelial tube. Such a protrusion is therefore constantly asso- 

 ciated w^ith those modes of vascularisation which Billroth calls 

 secondaiy and tertiary. It occurs most manifestly in the tertiari/ 

 mode. In this variety caecal protrusions may readily be seen to 

 occur at certain points in such capillary vessels as are already 

 pervious : these gradually extend, and finally bend round in the 

 form of an arch, to rejoin the parent vessel ; or two such protru- 

 sions may meet each other, with the like result of producing a 

 new capillary loop by their union. A more minute examination 

 demonstrates the presence of certain fine, threadlike processes, 

 whose exact relations are rather hard to determine, in connexion 

 wdth these ca?cal protrusions (fig. 33). The best object for this 

 investigation is undoubtedly the transparent border of the tail of 

 a tadpole ; but even here the vessels are surrounded by a number 

 of stellate connective-tissue corpuscles. Should we find the pro- 

 trusions from the capillaries continued into these corpuscles, they 

 may be viewed as identical with the threadlike processes, and we 

 infer that the development of the capillaries takes place by a 

 direct metamorphosis of corpuscular elements; should we fail, 

 on the other hand, in demonstrating any such continuity, the 



