382 DISCOVERY REPORTS 



either side of a muscle ridge and, though sometimes remaining connected by tenuous tubes of nuclear 

 membrane, they frequently separate completely. This phenomenon is to be regarded as a cytological 

 accident rather than as a regular, significant process that could be dignified by the title 'amitosis'. 

 Chromosomes lodge in the troughs between muscle-fibres and fail to reach the spindle at metaphase ; 

 the daughters of such divisions are therefore deficient. Chromosomes break transversely but, in some 

 cases at least, the fragments take their place on the spindle and divide normally. This is interesting as 

 it suggests that the spindle attachment is not a single, localized one, but is of the diffuse or polycentric 

 type. There is evidence too that the nucleoli, like the nuclei, can be squeezed out and caused to divide 

 if subjected to exceptional mechanical stress. 



The occurrence of mitosis in the nuclear layer has no visible effect on the underlying fibre layer of 

 the epithelium. This is also the case in the Hydra ectoderm, although in the endoderm the fibres are 

 said to be absorbed during mitosis and resecreted following its completion (McConnell, 1932). 

 Electron-microscope studies by Hess, Cohen and Robson (1957) show that in Hydra (both in the 

 ectoderm and endoderm) the muscle fibres are contained within basal outgrowths of the epithelial cells 

 and that the fibres do not fuse with one another, although they may come into close proximity. There 

 is no evidence therefore in Hydra for true anastomosis in the fibre sheet ; and the fact that fibres are 

 not affected by the mitosis of their nuclei cannot serve as a demonstration of any breakdown in the 

 simple cell-fibre relationship. 



In Physalia the fibres form what at first sight appears to be an anastomosing network. Individual 

 fibres can be traced for distances of over 300 fi without interconnections. Few fibres end freely; they 

 may vary in thickness, becoming thin and inconspicuous, but usually they can be traced back into the 

 'net' again. Where two fibres come together, the appearance is one of direct fusion (PI. XXVI, fig. 5). 

 In actual fact true anastomosis (that is, direct confluence of fibre material) may never occur. Instances 

 have been found where what appears to be a single fibre in the net has become bent or buckled at some 

 point, revealing itself as two closely applied fibres (PI. XXVI, fig. 6). The 'anastomosing net' may 

 be no more than an array of discrete but closely juxtaposed fibres, as in Hydra. 



Although in many preparations the nuclear layer appears syncytial, staining with iron haematoxylin 

 invariably shows up the cellular boundaries. A typical cell of the saccus-endoderm may have ten or 

 fifteen fibres running under it, some of which can be followed beneath six or eight other cells. 

 Attempts to relate particular fibres to particular cells have not been successful, even in regions where 

 the fibres are strongly contracted (PI. XXVI, fig. 7). Although we cannot at this stage determine in 

 any case which fibres ' belong ' to particular cells, we have no definite grounds for abandoning the 

 classical view of the muscle sheet as divisible into territories referable to individual cells. 



4. The nervous system 

 Chun (1882) was the first to identify nerves in Physalia, in the ectoderm of the gastrozooids. In the 

 related Rhizophysa, but not in Physalia itself, Chun discovered nerves in the float. With the exception 

 of Parker (1932) whose studies revealed that conduction in the tentacles is neuromuscular, not simply 

 muscular, no other workers on Physalia refer to the nervous system. Nerves have never been identi- 

 fied histologically in the tentacles. 



An attempt to stain the nerves in Physalia was made in the present investigation. It is possible to 

 make out the general distribution of the nerves in lightly stained carmalum and haemalum prepara- 

 tions, and Sudan-black staining of Ca-formaldehyde material sometimes colours the fibres fairly well. 

 However, the author's main efforts have been directed towards making silver preparations, since these, 

 when successful, are much the most revealing. Holmes's method (1947), which has been used 



