96
Geological Survey of Finland, Bulletin 395
Tapani Mutanen
of plagioclase, the most common mineral in
mafic layered intrusions; on the other hand it
was purported that the plagioclase crystals did
not settle as cumulus crystals but grew from
the bottom (see Campbell, 1978), as, indeed,
they sometimes do.
The news of the death of the cumulus theory
soon proved exaggerated, however, as geolo-
gists were reminded of the facts everyone al-
ready knew (see e.g., Morse, 1969a; Cox,
1985; Cox & Mitchell, 1988; Ghiorso & Car-
michael, 1985; Marsh & Maxey, 1985; Martin
& Nokes, 1988, 1989; Martin, 1989; Martin &
Nokes 1989). Most geologists involved with
the study of layered intrusions probably re-
garded the new dogma unworthy of a fight,
and paid no heed to it. Of one of the many fal-
lacies of the un-cumulus theory Morse stated,
rightfully: “The notion that the [cumulus] the-
ory deals only with crystal settling and molec-
ular diffusion is a recent myth” (Morse, 1985).
The other fashion, also at variance with clas-
sical cumulus models, was the concept of dou-
ble diffusive convection (DDC). This is essen-
tially a single phase compositional convection
model (McBirney & Noyes, 1979; Chen &
Turner, 1980: Tait et al, 1984), relevant in
low-viscosity fluids, such as water. Applied to
intrusions, it needs a calm magma and does not
tolerate overall convection (see e.g., Morse,
1985) and probably no stirring other than that
caused by the DDC itself. The DDC denies the
significance of crystal settling in magmas (Mc-
Birney & Noyes, 1979) and is, thus, another
aspect of un-cumulus thinking.
The third trend in explaining the filling and
crystallization of layered intrusion is what I
call the MMP (for multiple magma pulses) hy-
pothesis and its kin. Considering the huge vol-
umes of many layered intrusions, the incre-
mental intrusion of magma appeals to our com-
mon sense. Incredible as they may seem, sin-
gle eruptions of basaltic magma with volumes
of up to 6000 km
3
are known (McDougall, 1962;
Feoktistov, 1976; Williams & McBirney, 1979).
Actually, MMP is not a recent fad; proposals
of new magma surges can be traced back to
Kuschke (1939). When one sees a radical
change (a reversal or the appearance of a Dop-
pelgänger phase) in the cumulate succession, it
is only natural to invoke a new kind of magma.
It is seldom thought that this new magma
(mixture of liquid and crystals) might have
been generated in the magma chamber itself,
with no help from the mantle. Here I propagate
the idea that the new magmas were products of
“inside” events resulting from magma/wall
rock interaction.
In many well-studied intrusions the possibil-
ity of MMP has been considered but rejected,
and with good reason. In Stillwater, Hess
(1960), Jackson (1961) and recently Loferski
and co-authors (1990) concluded that the mag-
ma chamber was filled about to its final vol-
ume before significant crystal accumulation,
and the “oscillations” (layering, reversals)
were due to mechanisms operating inside the
magma chamber (Hess, 1960). There is a con-
tinuous iron-enrichment trend through the fa-
mous J-M reef, from footwall to the hanging
wall, with no evidence for a new magma pulse
at that important reversal (Bow et al., 1981). A
variant of the MMP hypothesis maintains that
a new pulse of a heavy anorthositic magma
was injected beneath a buoyant old magma
evolved from an ultramafic magma (Irvine et
al., 1982). The Pb isotope studies indicate very
convincingly that the Stillwater magma cham-
ber was filled to capacity before any essential
cumulus sedimentation (Wooden et al., 1991).
At Bushveld the old ideas of MMP have
been redesigned from time to time (e.g., von
Gruenewaldt, 1979; Sharpe, 1981). How was it
then that the old field geologists were not able
to see the MMPs? Cameron and Emerson
(1959) did not find MMP as satisfactory expla-
nation for the innumerable reversals, nor did
Cameron (1969, 1970) find signs of MMP in
the Transition and Critical Zones, the favoured
stratigraphic interval for MMP. Cameron and
Desborough (1969) did not find any breaks in
the px-pl compositional trends at ore intervals.