396 JOHN LEWIS BREMER 



more widely separated. No more branches are produced. With 

 this there occurs, apparently, an absorption of the peripheral 

 free ends into the network, leaving the outer set of cross connec- 

 tions as a series of arches, joining the ends of the radial cords (fig. 

 6). Also, since the cross connections do not lengthen so much 

 as the radial cords, the network assumes a distinctly radial appear- 

 ance. 



There now occurs a partial destruction of the network. Al- 

 though there is a general increase in the diameter of the cords, 

 certain ones, usually, but not exclusively, those forming cross 

 connections, remain of their original size or even become smaller. 

 Many of these attenuated connections soon become severed, strand 

 after strand, and the loose ends are absorbed into the network. 

 This partial destruction of the network goes on for a long time, 

 in the human embryo certainly from 22.8 mm. to 9.1 cm., perhaps 

 longer; in the later stages, when the cords have become more es- 

 tablished, the loose ends are not usually retracted or absorbed, 

 but remain as short knobs or as long branches with blind ends. 



The results of this process may be seen by comparing the models 

 of tubules from embryos of 37.0 mm. and 9.1 cm. (figs. 6, 7, 8 

 and 9), a full description of which will be given later. At a glance, 

 the destruction of the network and the consequent isolation of 

 certain cords can be easily traced, as the figures are still uncom- 

 plicated by convolutions. 



During this time the inner or central ends of the radial cords 

 have come into contact with the rete cords, which also have formed 

 a network. The rete network, the 'Keimdriisennetz' of Mihal- 

 kowicz and other older writers, is quite irregular, of small mesh, 

 and persists throughout life. It occupies the mediastinum testis, 

 and in the lower two-thirds of this spreads out in a fanshape, fill- 

 ing the space enclosed by the mass of the testis cords, which is 

 itself cresentic in section. A single testis cord may come di- 

 rectly in contact and join with the rete network, or peripheral rete 

 cords may extend far into the area of the testis network, so that 

 the boundary line between rete network and testis network is 

 irregular and wavy (fig. 6). These extensions probably indicate 

 the position of the septa of the adult testis, along which the tubuli 



