SHAPE OF THE NUCLEUS 597 
be as in Text-fig. 9, B on p. 600, which is never the case 
in nature. 
(2) That there is mutual repulsion between 
cell membrane and nuclear membrane. Such a 
force, acting in an inverse ratio to the distances between the 
two membranes is indicated in Diagram C, p. 600. This 
supposition explains : 
(a) Why nuclear and cell membranes practically never come 
into contact with one another. 
(b) Why the nucleus tends to elongate concurrently with 
the cytoplasm. 
(c) Why the nucleus is never round so long as the length of 
a cell is greater than its breadth, although there is often ample 
room in the cytoplasm for it to become spherical. 
Of the nature of such a force responsible for the antagonism 
between cell membrane and nuclear membrane we know 
nothing. 
6. CANALICULI IN THE NUCLEAR MEMBRANE. 
Intranuclear canaliculi are more common in spherical and 
oval nuclei than is usually thought. They have been described 
in the spermatogonia of Amphibia by Champy (4), and are 
easy to demonstrate in Rana esculenta andthe Axolotl. 
Canaliculi in the nuclear membrane occur in many types of 
cell; we have observed them in the epithelium lining the 
Wolffian duct in the salamander, and in pyramidal cells of 
the cerebral cortex in the guinea-pig. ‘These structures are 
illustrated in Pl. 24, fig. 4, and in Text-fig. 8. 
The intranuclear canaliculus is essentially a narrow invagina- 
tion of the nuclear membrane. Its blind extremity, which 
may be bifid, often ends in the vicinity of the nucleolus. That 
this structure is a definite tube and not a deep furrow in the 
nuclear membrane, is shown in transverse sections of it. In 
many spermatogonia there seems to be some relation between 
the canaliculus and the centrosome ; at the prophase the latter 
comes to lie very close to the former, often exactly opposite 
its aperture in the nuclear membrane. 
