THE MARSUPIUM OF THE UNIONID.E. 



33 



should di-appear after spawning. It is true that one occasionally 

 encounter- a partial fu-ion of two adjacent interlamellar junc- 

 tions, with a consequent division into two or more parts of the 

 water-tube lyiiiL; between them, but this is not constant in occur- 

 rence for the species and, when it is pre-ent at all, it involves 

 only a -iiiije tube here and there in the i;ill. \Ye have observed 

 such fu-ions in a few indi\idual- 

 bclon^inu to different genera in 

 both i;ra\ id and non-gravid gill>, 

 but it : ndition that mu-t be 



rded nicrch a- an occa-ioiutl 

 variation and i- entirely different 

 from that \\hich i- -uppo-<-d by 

 ( )rtmami t" e\i-i in the Anodon- 

 tin.e. Hi- de-i riiition. m 

 i- at total variance \\ ith our ob- 



-er\ ationa in the th- -. re- 



ferred t' >, a- -< lion- of the ^ill- 

 in the-e form-, taken at \arioii- 



\ id period, 

 -hov. :/</</ v no : 



diri.^inn nf the tu't'f s. h i ;. 



which are drawn from hori/ontal 

 section- of the v;ra\id mar-u|>ium 

 of Anoilontti cnttiniitii Sa\ . . 

 mitlnntti truniUi\\'nv.\}( and v 

 phynota compfanata Mann--. 

 ti\el\ , the water-tubes, containing 



embr\o- and :<lochidia, are -een in 

 their u-ual form, undivided lon- 

 gitudinally and bounded by the witli nlocliidia. x 31.5 Kline del. 

 inner and outer lamella- <i t In- ^ill 



and by he interlamellar junctions. If such a divi-ion of the 

 tube- into three part-, a- Uninann describe-, were present, it 

 \\onld of cour-e be indicated in the-e sections. 



\\ e are at a lo-- t,. under-tand what appearances observed by 

 Urtmann could ha\e i^i\en ri-e to his error. The only tiling 

 that Miji.ne>t> it-elf i- that the material which he used had been 



i complttnata 



', . \<>n of por- 



ti.ii. ipitini, 



a w.itrr-tulir. uiKlividciJ and 



