Priority effects

1
Priority effects
What are the consequences of phenological
patterns for interspecific interactions? Harper
(1961) planted two species of grass Bromus
rigidus and B. madritensis either simultaneously
or with B. rigidus sown 3 weeks after B.
madritensis Grown together B.rigidus accounted
for 75 of biomass Sown later B.rigidus
accounted for 10 of biomass What might
determine a priority effect like this??
2
Benke (1978) Priority effects on dragonflies -
Different dragonfly species have predictable
seasons of emergence from larvae to adults
early species - emerge synchronously in early
spring and complete breeding by mid summer
late species - emerge non-synchronously mid
summer. These spp oviposit when hatchlings of the
early spp are already swimming around the
pond Set up screened pens in ponds to control
when female dragonflies can oviposit. Showed
that early species significantly depress the
abundance of late species relative to
experimental cages where early spp excluded.
3
Treatment effects on abundance of late species
800
- early species
600
Number late sppper m2
early species
400
200
0
M
A
M
J
J
A
S
L abundance of early species reduced E
abundance of late species reduced EL open to
early and late species NO No dragonflies (some
did get in!)
4
Priority effects that arent tied to phenology
Schulman (1983) looked at recruitment of marine
reef fish from the larval stage on newly created
artificial reefs composed of concrete building
blocks - Recruitment of fish was inhibited by
prior occupation by two species of beaugregory
(territorial damselfish) and juvenile snapper -
New territories on the reef open at random.
Species that have more settling larvae available
when the territory opens up will have a higher
probability of filling it. Other examples??
5
Remember Drosophila competing for fruits? Very
similar study of Drosophila on mushrooms
(Shorrocks 1994) Probably many similar examples
of competition-colonization trade-offs are
relevant here Many examples in
succession Early successional species are poor
competitiors but arrive first at open habitats
because of superior dispersal Many examples of
priority effects in forest succession that we
will look at in coming weeks
6
Priority effects are an example of an Assembly
Rule
Jared Diamond (1975) first explored the idea that
there are rules that govern how communities are
assembled.
Diamonds work based on an accumulation of
observational data on the distribution of bird
species on Bismark islands around New
Guinea Interested to know if certain sets of spp
that could be drawn at random from a species pool
fail to coexist at some local level
7
Bismarck Archipelago
Islands N and E of Papua New Guinea
8
Diamonds approach to examining coexistence
Describes incidence functions the probability
that a particular species will occur in a
particular community given some attribute of the
community Predictive attribute of the community
in this case species richness
Diamond was clearly influenced by MacArthur and
his warblers - species fit together in a
complementary way in communities dictated by the
strength of interspecific competition
9
Species called High-S require more specialized
features of communities that support a variety of
other species. Tramp species occur islands
including those with low sp richness
Tramp sp
High-S sp
(dove)
(cuckoo)
Probability of occurrence
Tramp sp
Supertramp
(flower pecker)
(cuckoo-dove)
Species richness
10
Diamonds explanation for these
patterns Species use/consume resources (eg
food, nesting sites) out of the total resource
pool available on the island. Resource pool
determined mostly by island size? To determine
if species can coexist Subtract individual spp.
resource use from resource pool to see what
resources remain. Then determine if additional
species would be able to survive on those
resources Matching resource use and production
allows predictions to be made of allowed and
forbidden species combinations
11
4 spp guild with different resource requirements.
Solid line represents resource production. Dashed
line resource use by each species
Small island - only sp 3 exists
Larger island spp 2 and 4 but not 2,3,4 can
coexist
Larger island, sp 1 could invade island
occupied by spp 2 and 4
12
Diamond codified the patterns he observed into a
set of Assembly rules 1. Considering all
combinations that could be found for a group of
related spp. only certain ones exist in
nature 2. Those permissible combinations resist
invaders that would transform them to forbidden
combinations 5. Checkerboard rule - some pairs
of species never coexist either by themselves or
as part of a larger combination How would you
test if these assembly rules actually operate??
13
Example of Diamonds rule that some spp pairs
never coexist
14
Various tests of Diamonds rules using null
models Connor and Simberloff (1979) and other
papers - looked at whether fewer species
combinations occurred in nature than expected at
random. Could not reject the null model
Gotelli and McCabe (2002) - more complete
analysis of particular assembly rules - first
test of the checkerboard assembly
rule -Assembled data from 96 studies of species
occurrences from scales of 1-1010 m2 used Monte
Carlo randomizations to examine whether there are
species co-occurrences that are less likely than
expected at random -Found general SUPPORT for
assembly rules
15
BUT REMEMBER Evaluating the null model does NOT
test whether competition is responsible for
patterns of species occurrence Null models are
statistical tools for recognized non-random
patterns, not a litmus test for competitive
effects What other factor might result in
less-than-random patterns of species
co-occurrence?? Where have we seen good
evidence for interspecific competition before??
16

• MCloskey (1978) looked for the existence of
  assembly rules in communities of seed eating
  rodents in Sonoran Desert.
• - Collected detailed natural history on habitat
  niche and feeding niche of four rodent spp.
  Quantified differences in habitat characteristics
  (shrub density) and sizes of seeds consumed by
  each species and combined this information into a
  measure of niche overlap
• Found that observed pairs and triplets of
  coexisting rodent species were those that had
  both the smallest niche separation and the
  maximum resource utilization for a habitat
• cf predictions of Diamond - but resources
  strongly limiting in Deserts

17
M, P, A, and B are rodent species. Only
spp combos with Low Niche separation found.
Values in bracketsoccurrences
18
Fox (1987) expanded on MCloskeys idea If
competition determines which species can enter a
developing community, then the outcome of
community assembly should be predictable
according to the functional groups each species
belongs to. - Dont need to know all MCloskeys
details about what each sp eats exactly etc just
need to know how to assign it to a group (based
on taxonomy or ecology) itself not exactly
easy Foxs assembly rule there is a much
higher probability that each species entering a
local community will be drawn from a different
functional group until each group is represented
before the cycle repeats
19
Rule is based primarily on competition for food
if some functional group becomes
disproportionately represented in a local
community then competition will lower the
probability that the next spp will belong to that
group Fox called a community a favoured state
whenever all pairs of functional groups have the
same number of species in them, or differ by only
1 sp. Imagine a hypothetical site and 3
functional groups. Favoured states include
(2,2,2), (2,3,2) and (1,1,0). Unfavoured
states include (2,1,4) and (1,0,2)
20
If Foxs rule operates then should find more
favoured states than the random expectation.
Provides a null hypothesis that species enter a
community independently of their functional
group. Two data sets Nevada Test Site - 115
rodent communities. 92 found to be in favoured
states. Random-draw simulations mean expected
number 62.5. Probability of finding the
observed number lt0.001 Arid Southwest - 202
rodent communities. 128 in favoured states.
Random expectation 111.3 Probablity of finding
observed number lt0.01 Evidence therefore for
deterministic processes?? Maybe not...
21
Simberloff et al (1999) criticised the data
- Nevada Test Site was in fact used to detonate
a series of nuclear explosions 4 yr before the
study disturbance might be important?? -
Criticized the randomization procedure - e.g.
doesnt take into account the frequency of
occurrence of each species or its range
distribution. Reanalysis of Foxs data with
different assumptions in the null model failed to
find unexpectedly high numbers of favored states
(Stone et al. 1996) - Criticized functional
group assignments. Stone et al. randomized
species assignment to functional group.
Randomized species did not generate any more
favored states than did the biologically based
assignment.
22
Application of assembly rules to
restoration/invasion ecology 1) Are the
component species of a community mutually
selected from a larger species pool so as to
fit one another? 2) Does the resulting
community resist invasion - if so why? 3) Do
invasive species alter/dissassemble communities
(eg affect competitive interactions between
remaining species? 4) To what extent is the
final community specified by the physical
environment as opposed to chance
colonization? e.g. test how species arrival
sequence in successional communities influences
species composition.
23
Seems to be little formal design and analysis of
restoration projects in the context of ecological
rules -Goals for land managers (pattern,
outcome, specifics) and ecologists (process,
mechanism, generality) are different? -Need to
focus more on traits rather than on species How
are traits associated with environmental
conditions (e.g. defensive characteristics of
plants, structural characteristics, response to
disturbance).
24

• Conclusions
• Assembly rules idea not broadly accepted among
  ecologists except in a broad sense. Reflects move
  away from focusing on species interactions in
  determining community composition
• e.g. dispersal assembly vs. niche assembly
• Very difficult to test for assembly rules based
  on spp presence-absence patterns (too many
  explanatory variables that need to be ruled out)
• Even if patterns are consistent with competitive
  effects (Diamonds rules), other factors may also
  explain the pattern (range distributions,
  physiological tolerances etc)