STRUCTURE AND LIFE-HISTORY OF CRITHIDIA MELOPHACJIA. 203 



and a rosette (PI. 12, fig. 4) is produced, but the rapidity of 

 the process of formation of flagella causes short duration of 

 the rosette stage. On the other hand, repeated longitudinal 

 division of flagellated individuals occurs, and as the individuals 

 so produced do not separate immediately, rosettes (PI. 12, 

 fig. 56) ai-e formed. In division, the posterior ends of the 

 daughter-organisms are the last parts to separate. As the 

 daughter-forms remain in proximity and themselves proceed 

 to divide with rapidity, true division-rosettes ai-e formed, in 

 which the posterior ends of the organisms are central, while 

 the flao-ella radiate out from the common centre. Such 

 division-rosettes (PL 12, fig. 56) differ from the aggregation- 

 rosettes (PI. 12, figs. 41, 43; PI. 13, figs. 95, 96) where the 

 organisms become attached by their flagella. The distinction 

 between the two forms of rosettes has not been shown by 

 previous workers on C. melophagia. 



Longitudinal division results in the formation of both 

 equal and sub-equal daughter forms. 



While the occurrence of equal longitudinal fission is the 

 commoner (PI. 12, figs. 50, 54, 55), I have seen cases of 

 marked inequality in the size of the daughter-pai^asites, the 

 one being very thin and narroAv, the other considerably 

 broader and thicker (PL 12, figs. 51, 53). As the entire 

 process of sub-equal division has been watched in living' 

 organisms, there is no possibility of it being mistaken for 

 anything else. The polymorphism resultant on division is 

 strongly against the idea that there are sexual forms of 

 Crithidia, and I have never seen the slightest indication 

 that there is sexual dimorphism, in C. melophagia, 0. 

 gerridis, Herpetomonas jaculum, H. muscEe domes- 

 ticae, H. culicis, and a new Herpetomonas from Vespa 

 crabro, all of which I have examined in the living con- 

 dition (see Porter [1909] on C. gerridis and H. jaculum). 



Division, usually twice repeated, is found to occur in 

 parasites destined to encyst, and the resultant forms are very 

 small. The first division is of the usual flagellate type 

 (PL 13, fig. 97). The process of the second division rather 



