300 
lilt' Aliuriril II ( trill (HI isf . 
'I'lic pnpcr is t:iU('ii up iii:iiiily with tiic iccord of I'licts (Icsci'ip- 
ti\ r (if tln' iiiicio-stiiicliin' of tlic crystMlli/.iilioiis seen in tln'sc 
splu'iiilitcs. Tlicic is. however. ;i liypotlicticnl iiiidcilyiii^ cmisc. 
<)!• coiu'cptioii. whicii .Mr. Iddiiiu's cntci'tniiis ;is the iuitiiil coikU- 
tioii pri'C('(lfiit for the foruuitiou of >.phcnilit('s. viz. : the pres- 
t'lU'c :m(l uii('((ii:il distrilmtioii of \v:itcr-\ iipor witiiiii the siliceous 
ni:i<ilii;i. 'Phis w:itei'-V;ipor seems to net. :is })resniiled \\\ the 
aiitliof. r;ither pliysicMJly lii:m eheiiiie:dly :iiid to deleniiiiie hy 
its i>r(.;iter iimouiit in loe:dities tiie <>re;iter inoh'cnlar inol»ility 
within the molten mauinn which locally permits eryst:dliz:ition. In 
otiier words the coiice|)tion of tlu' author is that wliei-e now exists 
a spherulite within the olisidian glass, oi-iiiinally was a local cen- 
tre of ui-eater hydration than in the mauina in <>-eneral. This con- 
dition precedent answers, in .^il■. Iddings' hypothesis to the col- 
loidal <L>'lol»ule in Mr. C'rosss. lint neithei' of the hypotheses aims 
to account for the initiation of the actual crystallization. 
.\t tii'st uianci" there seems to lie an olistacle to this assumetl 
nne(iual distribution of hytlrie vapor, and it is the same in charac- 
ter, as that stated liy >Ir. Iddiims a*>ainst an une(|ual distriliution 
of heat • •througliout any considcraMe mass of the nniunia" (|). 
44(>) viz.: /// fill iiiitmi i)f fill- iiisi , |le;it and hydrie \apor. in tiieii' 
manner and deiifee of distribution, ai'e. when l>oth are present, if 
not identical, so intimati'ly correlati\e that what may lie assumed 
for one. may. pei'liaps. !»e assumed for tlu' other. IJut it must be 
admitteil t hat t here is a fundamental distinction to be ol»served. 
in this casi'. between the presence antl pi'olialile mannei- of distri- 
bution of heat and the presence and piobable distriliution <»f 
water- vapoi'. lleat. in a molten eru[ili\i' rock, is pi-imary and es- 
seutinl and may reasonably be supposed not to present ^reat var- 
iations throughout ••any considerable mass. " liut hydrie vapor is 
secondary and accidental, and may be siip[»osed to be most 
abundant at those jioints where the molten mauina may have Ih'cu 
br<»u<:ht into contact with superlicial waters. Such acciilental 
contacts w<iidd. in the lirst place, cause a rapid and general diffu- 
sion of water-vajior through all the adjoining crust, as well as 
through the molten magma, and it wouhl seem inevitalile that 
some pai'ts of the magma, even throughout •considerable masses' 
Avouhl be atfecteij dilfereiitly liy its presence in ditfering amounts. 
Notwithstanding this distinction, however, in the nature of tiie 
origination of these lw<j I'li'ments in the licpiidity of the magma, it 
