2/6 JERI.MIAH S. FERGUSON. 



connective tissue possess locomotion and move largely in one 

 direction, but their motion and locomotion is extremely active, 

 and so far as I can observe in the living animal they leave behind 

 no trail of secretion, certainly no observable granules: still more 

 important, the fibers appear to have been formed in great numbers 

 prior to the appearance of the type of spindle cells in any con- 

 siderable proportion relative to the other types of cells, round 

 and stellate, already present in the primitive connective tissue. 

 Hence I cannot conceive of the spindle cells as producing fibers 

 by direct secretion. 



But it is at the time of the predominance of stellate cells that 

 the fibers make their first appearance in the fish's fin, as is 

 likewise the case in the tissues studied by other observers. 

 These stellate cells rapidly change their form, throwing out and 

 retracting their processes in rapid succession, as can be seen in 

 any living, free-swimming /'"nmlnlKs embryo under 25 mm. total 

 length. Fig. I shows the outline of such an early connective 

 tissue cell in the caudal fin, its changes of form at intervals for a 

 period of twenty-four minutes being accurately depicted with the 

 aid of the camera lucida. It is impossible to both watch and 

 draw all the changes in form occurring during this period, but 

 Mifficient are shown to demonstrate very active amceboid motion. 

 By observing these cells in relation to a relatively fixed object, 

 e. g., a chromatophore, or a joint in an adjacent fin-ray, I have 

 been able to detect in them considerable locomotion. 1 find 

 that, unlike the spindle cells, the stellate cells appear to travel 

 equally well in all directions. Neither following their locomotion 

 nor in trail of a retracted process do I find any indication of 

 secreted granules, nor of anything other than the most delicate, 

 preformed fibers. One would suspect that it new fibers were 

 being formed by a process of secretion they would appear within 

 i \\enty-foiir minutes after such a cell or its process had occupied 

 a gi\en position in which liber should reasonably be expected 

 to be formed. Though I ha\ e repeated the experiment many 

 times I have failed to olMT\e such evidence of secretion, and 

 the-e re-nlt-, do not appear to be con^i-tent with a theory of 

 dire> i -eiTetion of libers by the primitive connective tissue cells. 



The intr.ii t llnlar theories of the origin of connecti\e li^-ue 

 fibers may be divided into I wo classes, those in which the libers 



