THE SPERMATOGENESLS OF DOMESTIO JIA-MUALS. • 241 



in the guiucn-pip. Ho st.ites : " Lcs oljsorvatious, oucdvo nii-es (Vaillomg, sur 

 Ics clioudriüsomes des oellixles SL-miualos dos ^^cl•tel>lvs, luonh-uiit (iiie, si dims 

 lea clnsses iuforiciires (Aiuphiliiijus) il existe des clujiifli-i(Ä)moa de forme 

 tiliimentciiso, In fovino gi-iinuleiise patiit eh-e la. regie cliez Ls Mamuii- 

 ft^'n«. G^tte fonue so iiiaiutieut ^jeudaiit les divisions de matimitiou : ou ohservc 

 lion, jK^diUit la soodudo division do col)aye, quelques choudiüouiitos, niaLs la 

 foriualiou do t-hondriomntes reguliers, comnie cliez les lusectes, n'a ])ag encore 

 t'to sigualec jnsqu'ifi. Dans la sjjenmitide des IVL^nnmferes, les iiiitoclioudi'ies 

 no so condcnsont ytus en eori« liomogoiie, comparable an Nelioukcrn, nmis 

 rcstont indi\idiuilisfes. Tons les oljsoiTateurs sout d'aecord sur lour sort final : 

 ellcs formeut nue gaine, do strnctiu-e variable, an filament asile du spermiitc>- 

 zoide." As to the process of the aiTaugemeut of the mitocliondiial granules 

 wound the axial filament of the spermatozoa, contraiT to mv observation, he 

 states that '• lo processus de la formation do la gaine mitochondiiale debute 

 \)nx consequent, chez le cobaye comme chez le rat, an voisinage de la tete." 

 In my materials, on the other liaud, tlie mitochondiial granules begin to 

 aiTango around the axial filament near the jrsterior coll wall. 



With regard to the origin of the mitochondrial giMUules tlie opinions of 

 the investign.toi's do not agree, some (Goltlschmidt, Buclmer, Jordan) taking 

 the \-iew tliat these granules ure formed by extnision fivnu the nucleus, while 

 others (Moves, Duesberg) hold for the cv-toplasmic origin. Jordan ('12; in 

 o^KSSum, though could not directly observe the extrasion of the granules 

 thi'ough the nuclear memljraue, concluded that the mitochondrial gi'auules 

 originate fi'om the nucleus. This is based upon the following facts : 1. Within 

 and without the nucleus nianj' granules appear which have a simiLxr form 

 and stiiining capacity to those of the mitochondiüal granules. 2. lu thess 

 cells one half of the periphery of the nuclem becomes surroimded by a compact 

 mass of spherical or bivalent clu'omidia closely adhering to the nuclear wall. 

 3. The cell in this condition Ls cliaracterised l)y the absence of chi-omatin 

 particles witliin the nucleus. 4. Tlie appearance of the chondriosomes in the 

 cj-toplasm is c( .incident with the disapjjearance of the chromidia on the nuclear 

 nieml)rane. 



In my materials it Ls hard to make sme ^^■hether the mitochondiial 

 granules aiise from the cjiopUism or from the nucleus. Such gi-anules as 



