742 THOS. H. MONTGOMERY, JR. 
junction of the univalents. Fig. 74 shows two entire gemini and 
fig. 75, of another cell of the same cyst, three entire ones. Figs. 
76 to 80 show gemini from cells of one cyst, of which figs. 76 and 
78 show each two gemini and the others each one. Fig. 81, of a 
stage when the gemini are more condensed, shows four whole 
gemini; fig. 82, a cell of the same cyst, two whole gemini; and fig. 
83, still another cell of the same cyst, shows all the gemini. The 
last three figures, from cells of the same cyst, illustrate how 
rapidly these changes succeed each other. In figs. 84 and 85 
all the chromosomes are shown, and in each of these cases there 
are six bivalent autosomes.® 
By the stage of fig. 85 most of the gemini have become straight- 
ened out, though two stillretain the angular form. In the late 
diakinesis of fig. 86, in which again all the chromosomes are drawn, 
all the autosome gemini have become straight with an annular 
constriction about the middle, the ‘dumbbell shape’ of my previ- 
ous descriptions. 
How are we to evaluate the definitive gemini of the first matura- 
tion mitosis, shown in figs. 87 to 89, and 92 to 94? Their history, 
as we have followed it, demonstrates that the twelve univalent 
autosomes were at first distinctly separate, they do not at any 
time produce a continuous spirem, they undergo a parallel con- 
jugation, then in each geminus the two components begin to 
separate at one end while remaining attached at the other (this 
the persisting conjugation point). Thus two parallel autosomes 
change into a V-shaped geminus, and finally the two arms of 
the V open out into a straight line. Accordingly, each geminus 
of the first maturation division represents two univalents placed 
end to end, and the constriction around each geminus marks the 
point of persisting conjugation of the two univalents. 
No other interpretation seems justified after along and repeated 
study of the whole series of phenomena. And a degree of cer- 
tainty can be reached because there is no rest stage at any period 
5 In fig. 84 one of these gemini has its two components, s, s, separated from each 
other. This is a frequent variation, and the history of such s-chromosomes I 
have detailed elsewhere (10a). 
