184 
George Arnold 
pairs bas beeu coimted in a large number of cells, and is approxi- 
mately 15, bnt as tbis stage is rapidly passed over, cells favourable 
for counting are not easily found. 
It will be seen from tbis, tbat in Hydro pbilus tbe reduction 
takes place before tbe syuapsis. Anotber uotewortby feature is tbat 
no spireme is formed, altbougb linin Strands connect up tbe cbro- 
mosomes witb eacb otber. 
Tbe synapsis is of long duration, and bas several remarkable 
features. 
An early stage of tbe same is seen in Fig. 11. Tbe gemini 
gradually mass up togetber in tbe centre of tbe uucleus. In no 
cells does tbis contractiou occur at oue side of tbe uucleus, and 
since tbere is no real spireme, tbis pbase in Hydropbilus bas a 
cbaracteristic appearance wbicb differs from tbat of uearly all otber 
animals, in wbicb a synaptic stage bas been sbown to occur. 
Wben all tbe gemini bave coutracted up to tbe centre of tbe 
nucleus, nearly all tbe linin tbreads bave disappeared (Figs. 12 and 
13). Tbey are probably witbdrawn into tbe synaptic aggregation, 
but tbis could not be definitely ascertaiued. 
Tbere is oue darkly staiuiug body, wbicb is never contracted 
up witb tbe otbers, (N Figs. 12—14), wbicb is tbe nucleolus. It is 
probable tbat it was separated out as early as tbe telopbase of tbe 
last spermatogonial divisiou, but as it takes tbe basic stains just as 
beavily as auy of tbe gemini, and possesses no distinctive sbape or 
cbaracter uutil tbe synaptic period is almost over, it is not possible 
to distinguisb it from auy of tbe otber deeply stained masses in tbe 
nucleus. Tbis body, in tbe earlier stages is connected witb tbe otber 
stained masses by linin tbreads (Fig. 12), and bas generally tbe sbape 
sbown in Fig. d 24 but occasionally eveu at tbis stage it assumes 
tbe sbape sbown in Fig. d’ 24, viz: a tbick uneven ring, very tbin 
or open at two places. 
Wben all tbe gemini bave eutered into tbe central contractiou 
(Fig. 14), tbere is seen a mass of cbromatic material in tbe centre 
of tbe uucleus, in wbicb no separate entities can be distinguisbed 
eveu in very ligbtly stained aniliue sections, or in sections stained 
witb Iron-alum-baematoxyliu, and snbjected to long extractiou. 
Long loops of linin now proceed out of tbis mass, but are not 
connected witb tbe nucleolus, wbicb by tbis time is free from all 
tbe otber nuclear contents, and lies on tbe uuclear membraue. 
