433 



exterior by a single canaliculus. However it is possible that the 

 canaliculi of fully formed lacunae may in some cases remain per- 

 manently as branches of the main canal from which they arose. 



At D large, elongated lacunae are seen, some of them constricted 

 transversely as though about to form several smaller lacunae, although 

 this appearance is not as common as the branched one described 

 above. At E are a few lacunae belonging to a trabecula running at 

 right angles to the plane of the paper. 



After careful study I have come to the conclusion that each 

 lacuna is connected with a single canaliculus, which in every case 

 observed sprang from the pointed end of a lacuna and, bending at 

 right angles, runs in a radial direction towards the periphery of the 

 trabecula. 



It is difficult to say how the cells get into the lacunae. They 

 do not migrate bodily into the axis of the trabecula from the 

 trabecular epithelium, neither have they been buried there by the 

 successive deposition over them of layers of chiten, for the main body 

 of the trabeculae is apparently laid down long before any lacunae 

 are visible in it. The canaliculi however appear to be present from 

 the outset. They probably contain some substance which has the 

 power to dissolve out the solid chiten thus making room for the 

 protoplasm of the lacuna cells. The nuclei must either be formed in 

 the protoplasm that collects in the growing lacunae, or else they 

 must migrate there from the surface epithelium through the exces- 

 sively small canaliculi. Either supposition seems improbable, never- 

 theless, the following observation gives some support to the latter 

 view. In the region of the olfactory organ, where there are no 

 lacunae, I have seen the pore-canals of the surface cuticula, when 

 stained in haematoxylin, filled with minute dots that looked like very 

 small, but sharply stained nuclei. Not then knowing of the existence 

 of the lacunae, I let the observation rest, hoping to find later some 

 explanation of the remarkable fact. 



A cross section of a trabecula shows the same general arrange- 

 ment of parts seen in a section of a single Haversian canal system. 

 In the axis of the trabecula is the dark yellow core surrounded by 

 the colorless, concentrically laminated chiten. In the yellow core is 

 a circle of lacunae, several rows deep, the largest and evidently the 

 oldest being innermost. In its centre is a small area devoid of 

 lacunae which, in some cases, is marked by an imperfect perforation, 

 or canal, running lengthwise of the trabecula. This canal, the walls 

 of which are closely opposed, is probably the remnant of what might 



