Kilauea Iki, 1959— Macdonald and Katsura 
367 
columns 4 and 5. Furthermore, if the conclusion 
that the pool of lava beneath the crust was 
largely liquid at the end of the eruption is cor- 
rect, the thinness of the solid crust (less than 
20 ft.) seems incompatible with degrees of 
crystallization of 90' or even 75 per cent. Also, 
in column 3 the pyroxene is considerably more 
calcic than would be expected to crystallize from 
a magma of the composition of the Pele’s hair. 
Columns 6 and 7 of Table 3 show the com- 
position of material that would have to be sep- 
arated from a parent magma of the average 
composition of analyses 2 and 3 (Table 1) to 
yield magma of the composition of sample 10, 
assuming about 25 and 50 per cent crystalliza- 
tion respectively. The mineral compositions in- 
dicated in the norms appear quite reasonable to 
have crystallized from magma of that composi- 
tion. This appears to furnish further corrobora- 
tion that samples 2 and 3 more nearly represent 
the parent magma of the drill-hole samples, and 
to indicate that sample 10 could have been de- 
rived from that parent wholly by crystal dif- 
ferentiation. 
In Figure 2, as in Figure 1, samples 2 to 8 lie 
along a trend controlled largely by movement of 
olivine. The sharp divergence of samples 9 and 
10 from that trend can be explained largely by 
the onset of more abundant crystallization of 
pyroxene. The position of the average tholeiitic 
pyroxene given by Kuno ( I960: 128) is plotted 
on the lower boundary of Figure 2, and it will 
be seen that a trend line from sample 8 through 
samples 9 and 10 intersects the silica scale close 
to the pyroxene point. It will also be noted that 
the norms of columns 6 and 7 (Table 3) con- 
tain considerably more pyroxene than does that 
of column 8, which represents the material that 
must be separated from the same parent magma 
to yield sample 8. The derivation of the alkali- 
rich sample 10 by increased crystallization of 
pyroxene agrees with Murata’s (I960) sugges- 
tion that crystallization of pyroxene is an im- 
portant factor in the formation of the alkalic 
basalts. 
Thus, it appears possible to derive all of the 
Kilauea Iki rocks by crystallization differentia- 
tion, and unquestionably this process has been 
of prime importance. This does not, however, 
prove that no other process has been involved. 
PERCENT SiO, 
Fig. 2. Alkali : silica diagram of Hawaiian lavas. 
The compositions of the lavas of the 1959 eruption 
of Kilauea Iki are shown by crosses, those of rocks 
belonging to the tholeiitic suite by solid dots, and 
those of rocks of the alkalic suite by open circles. The 
point for olivine represents phenocrysts of that min- 
eral in the 1959 lava; that for pyroxene is an average 
of tholeiitic pyroxenes listed by Kuno (I960: 128). 
The dashed line indicates the trend of differentiation 
of the Kilauea Iki drill-hole samples. F indicates the 
position of the iron-enriched segregation veinlet of 
Kilauea, and G that of the granophyre from Palolo 
Quarry (Kuno et al., 1957). 
Indeed, it would seem to be a foregone conclu- 
sion that other processes must have been going 
on in the magma and must have, to some de- 
gree, affected the composition of the magma. 
Such effects are most likely to have been ap- 
preciable in the rocks that depart from the main 
trend of differentiation, such as samples 9 and 
10. The departure of these samples from the 
trend results principally from their greater rich- 
ness in alkalies. The transfer and concentration 
of alkalies in one way or another as a part of 
magmatic differentiation has been suggested by 
many writers. It has thus far proved difficult to 
demonstrate to the satisfaction of petrologists in 
general, but this does not seem an adequate rea- 
son to omit it from consideration. 
