Agriculture Reference
In-Depth Information
One of the key differences between soil storage and canopy storage strategies is
resilience to unpredictable fire cycles. When the fire interval exceeds the life span
of the parent plant, soil storage is a viable strategy (e.g. Keeley
et al.
2005b
) but
canopy storage is not (Lamont
et al.
1991
; Enright
et al.
1998
; Lamont & Enright
2000
). Another important distinction is inherent differences in potential lifetime
reproductive output. Canopy seed storage is limited by determinant growth pat-
terns that restrict lifetime seed accumulation to the available canopy space,
whereas seed output is potentially much greater in species that disperse seeds for
soil storage. Seedbank studies for canopy-storage plants have shown that the
magnitude of seedbank storage is between 10 and 10
2
(occasionally 10
3
) seeds
per plant (Gill & McMahon
1988
; Cowling
et al.
1987
; Witkowski
et al.
1991
;
Lamont
et al.
1999
; Bradstock & Cohn
2002b
). Seedbank estimates are much more
difficult to obtain for soil-borne seeds and few have provided seed/parent ratios,
but in a single good year most are capable of producing between 10
3
and 10
4
seeds
per plant (Keeley
1977
; Luis-Calabuig
et al.
2000
). However, seed output and seed
storage are difficult to compare because competitive outcomes are not determined
solely by numbers but by seed size and seedling attributes. Nonetheless, postfire
seedling recruitment is often much greater for soil-storage species than for canopy-
storage species (
Table 9.2
).
The evolutionary forces selecting serotiny over soil storage may be diverse
(Lamont
et al.
1991
), but some hypothesized factors seem more likely than others.
For instance, it has been suggested that postfire serotinous seed release is a form of
mast seeding and this synchronized seed release may satiate postdispersal preda-
tors (O'Dowd & Gill
1984
). It seems likely that predation is an important driver
behind the high incidence of serotiny in South African and Australian MTC
ecosystems. Indeed, this pressure has been a factor selecting for the tough woody
cones in diverse lineages. Despite this extra investment in protective structures,
seed predators can destroy a large fraction of the canopy seedbank (Groom &
Lamont
1997
; Mezquida & Benkman
2004
).
Lamont and Enright (
2000
) hypothesized that greater reliability of precipitation
in Australian and South African MTC ecosystems was a factor selecting for
serotiny because seed germination is restricted to the year of seed release. In their
view less reliable rainfall as in California and the Mediterranean Basin would
favor soil-stored seeds, which could hold over to subsequent years if the first
postfire year were unfavorable. However, there is little empirical support for
this model as seed carryover after the first postfire year is largely inconsequential
in many soil-stored seedbanks. For example, in chaparral shrubs the first
postfire year recruitment typically comprises 95-100% of the total seedling
recruitment during early seral stages, even if rainfall is below average (see
Table 3.3
). In addition, if rainfall unpredictability were a critical selective factor,
there is nothing about the serotinous habit that would prevent gradual release over
multiple years (e.g. in the serotinous
Hesperocyparis
(
Cupressus
)
forbesii
second
postfire year seedling recruitment comprises 5-10% of the total, J.E. Keeley
unpublished data).