SUPERNUMERARY CHROMOSOMES IX CKUTHOPHII 221 



in the tclophase (Fig. 5). Figs. 6a and 6b are sister spermatocv 

 of the second order, showing X in a separate vesicle: this does 

 not happen in by any means one half of the second spermatocytes, 

 X l.rin:/. I ^-hould say, more often included within the same 

 nm Ic-ar mem! T.ui'' with tin- other chromosomes. 



1 i 7 is from the individual which had one supernumerary, 

 am! I )'<-. 8 to 14 from tin- < >ne that had two. 1 ilso >h-nv- 



a less ii-ual p"-iiioii of A", near the equatorial plate. As in the 

 /'/ 5l the supernumeraries mayor max- not 



divide in the t"ir-t maturation mitosis, and they may, when un- 

 di\i > eiiher pole of the spindle. Their position in the 



met.ipha--e does not nec.-^.irily indicate whether they will di\ id. 

 or in-!. 1 In- di'ti-niiiniiig factor is probably the attachmen; 

 -pindle ill MI OIK- n both poles. The supernumeraries 



in Figs. 7, >> and ') would not divide in this mitosis, but the 

 piv-en. i.iry in an equatorial plate or between 



daii;JiiiT pi. HI ~ d. - essarily assure its division. In I 



i" one i- di\i<l<-d. tlie other undivided, while in Fie-. 11 and u 

 both are divided. In the telophases shown in 1 i^-. i.> and 14, 

 both -np< rnumerai : divided in one case and neither in the 



othi-t. I 'hi p(--ible c i.mbinations of A' and the 2 -npernume- 

 raries in the ^pcrmaiid- are A", A" + is, X + 25, o, 15 and 

 loiiroi the possibilities are shown in Figs. 15 to [8. 



As to the origin of these supernumeraries, there is little evidence 

 in tin- material. In MctapodlHS Wilson ('09) disco\end the 

 pmbable i.ii^in of the -npeniumeraries in an irregular ^erond 

 >pei -maiiH -\ te miti.~i- in which both " idiochromosomes " \\ent 

 ! the -ami- IM.IC ot the >pindle, and therefore to the same -penna- 

 lid. The supernumeraries are thus shown \ be duplicate- ..i 

 the -mailer " idii .chromosome" in Mctapodius, or in one case of 

 .in " ":-chroMio-i niie" *io). In neither Diahrotica nor in Ccntlio- 

 philn* i- there a -mailer mate for the A' chromosome pre-ent. 

 The three supernumeraries which have been observed in (.'cntho- 

 philus are of about the same size; considerably less than mie half, 

 and app.irciitU about one fourth the size of A'. The behavior 

 of the -II|H iniimeraries in growth and re-t -tai^e- of the nucleus 

 indicate-; their probable relationship to A', and their behavior 

 in mito-i-. dividing only once, either in the lir-t or the -econd 



