TRICHOMONAD FLAGELLATES. 323 



homologue of the central spindle, each being a centrodesmus." This 

 homology to the central spindle is obviously a different thing from 

 its reference to the axostyle and is in nearer conformity to the correct 

 interpi-etation of this structure. Alexeieff (1914a) solves the prob- 

 lem by the purely objective term blepharoplastodesmose, objection- 

 able because of its length. 



A thread-like chromatic structure connecting recently divided 

 blepharoplasts or basal granules in Copromonas subtilis is figured by 

 Dobell (1908) without comment as to its significance. The bleph- 

 aroplast lies atijacent to the reservoir some distance anterior to 

 the nucleus and the paradesmose-like strand persisting between the 

 daughter blepharoplasts after their division, lies on the wall of the 

 vesicle much as the paradesmose of Trichomonas on its nuclear mem- 

 brane. It later disappear^ as in Trichomonas. We regard this 

 structure as homologous with the paradesmose of Trichomonas. 



As noted above our researches show that this chromatic thread 

 connects at first the dividing daughter blepharoplasts. When, 

 however, these part each into a centrosome and basal granule at the 

 poles of the spindle the paradesmose connects the basal granules, and 

 not the ccntrosomes. The paradesmose thus connects here those parts 

 of the two blepharoplasts which are attached to the extranuclear 

 chromatic apparatus and not that which stands at the poles of the 

 spindle. This relationship, added to the fact that it is at all times 

 extranuclear as well as outside of the spindle, raises some doubts in 

 our minds as to the precise homology of the paradesmose with the 

 central spindle. The name we propose avoids the length and partial 

 inapplicability of that proposed by Alexeiefl^ and is in part at least 

 descriptive of the relations of this structure. 



Since the instances of the division of the blepharoplast into centro- 

 some and basal granule seen by us are found mainly in Trichomonas 

 augusta and are few in number and since the division may represent 

 merely a precocious division of blepharoplasts, possibly followed by 

 re-fusion (PI. 2, Fig. 2.3, PI. 3, Figs. 24-2S), the resulting lack of a 

 precise homology of the paradesmose with the central spindle should 

 not be allowed to obscure its relationships to that structure. It is 

 in any event a differentiation of the central spindle modified by two 

 specializations in trichomonad mitosis, (1) the continuity of the 

 nuclear membrane which excludes it from the typical axial position 

 of the central spindle, and (2) the connection of the blepharoplast 

 with the extranuclear motor complex. 



The paradesmose persists for some time after the daughter nuclei 



