391 
Ren Bi bre a 
len or Sow rc | ee - 
ea = Ze een © 
| | 
Boe. | 51-86 | 3267 | 37-14 | 13:74 | 4641 | 4693 |= -52 
в... 33 21 21 18 03 
MO ON se 50 33 tlh 
EO | 18°35 | 11.56 | 3160. | 11-69: |, 23°25 | 23:59 |e 
Fe,0, 650 | 410 | | 10 | 337 | 73 
Е. | 368 | 2:32 | 2:32) 216 | ale 
MnO : DE DRE 15 08 | -07 
MgO .. | ‘08 "05 | | | ‘05 | trace "05 
DO... aie EN rt err T 1 В 
1.02... | 13°26 | 8-35 | 25:60 | 947 | 1782 17-40 "42 
iO... | 2-91 | 1:83 1:83 | 2:96 | = E13 
Beer... I 1:13 ai Tale US 42 295 Е 
| 100-25 | 63-16 | 101°65 | 37°61 100-77 | 100-67. 
strongly enriched in sodalite which was the first mineral to 
crystallize, and consequently it must have formed at the upper 
border of the magma chamber, where heat was being continu- 
ally lost by conduction. The layer enriched in sodalite would 
consolidate as naujaite, and the remaining magma would be 
left with the composition of lujavrite. As a matter of fact the 
naujaite does constitute the upper layer, and is underlain by 
the lujavrite. In its broad features, therefore, the mode of oc- 
currence of the rocks is in harmony with the hypothesis. Upon 
closer examination, however, it will be seen that fractional 
crystallization has probably not been the only cause of differen- 
tiation. For if it had, we should expect to find naujaite, not only 
aS an upper layer, but also as a border zone along the walls 
