Size Ratios

1
(No Transcript)
2
Size Ratios

• Much of the size ratio literature is based on
  simple univariate measurements of body size or a
  single character, and that may be misleading or
  simplistic (why?)
• For example, tests of morphological patterns
  frequently yield different results depending upon
  which character was used
• Additionally, most morphometrics are highly
  correlated

3
Size RatiosMultivariate Analyses

• Consequently, it may be more advantageous to use
  multivariate analyses of morphological pattern to
  gain insight into possible community structure

4
Size RatiosMultivariate Analyses

• There are a variety of multivariate methods
• For example, Principal Components reduces the
  dimensionality of the dataset and maximizes
  amount of variation explained or Discriminant
  Analysis (maximize separation between groups)

5
Size RatiosMultivariate Analyses

• Unfortunately a common practice is to represent
  species as points in multivariate space, drawing
  polygons (or amoebas) around groups of species
  and offering some post hoc explanation for the
  groupings

6
Size RatiosMultivariate Analyses

• However, there are still null models that can be
  generated to compare patterns of community
  structure from random patterns

7
Size RatiosMultivariate Analyses

• Two types of community structure have been
  inferred from multivariate analyses of morphology
• 1) overdispersion of morphology (assumed to
  reflect competition), analogous to size ratio
  analyses (measured within a community)
• 2) convergence of morphology (typically measure
  between communities)

8
Size Ratiosoverdispersion of morphology

• Gatz (1979) examined 56 morphological characters
  measured for co-occurring stream fishes
• Null assemblages were constructed by choosing
  random points along each factor axis and then
  projecting them into the morphological space
  (Euclidian distance very similar to RA1, where
  resource utilizations are replaced with random
  number and always get a relatively high)

9
Size Ratiosoverdispersion of morphology

• Of note, he also analyzed Euclidian distances
  between sympatrically occurring members of a
  single family or genus and obtained comparable
  results

10
Size Ratiosoverdispersion of morphology

• Ricklefs and Travis (1980) constructed 2 null
  models drawing species lists from a larger
  source pool and maintaining observed
  morphological features or randomly generating
  synthetic species by substituting each factor
  score with a random, normal deviate.
• Species packing was measured by the average
  nearest-neighbor distance and evenness as the
  standard deviation of this metric

11
Size Ratiosoverdispersion of morphology

• Applying this protocol to avian communities, mean
  nearest-neighbor distance were generally less
  than expected, whereas standard deviations
  usually matched the predictions of the null model
  (Fig. 6.10)

12
Results varied widely with island communities
being overdispersed
13
Size RatiosConcordance of Morphology

• These studies involve 2 assemblages of unrelated
  species in similar environments
• One option would be to test for ecological
  equivalent sets of species
• Null Model differences in body size of species
  matched between two assemblages are no smaller
  than would be expected by chance

14
Size RatiosConcordance of Morphology

• Size distributions were too close to be
  expected by chance
• However, concordance may not be best test for
  similarity

15
Size RatiosMorphology and Abundance

• Previous assumption morphology and resource use
  are intimately related and competition is so
  strong, cannot coexist
• Morphology and abundance of coexisting species
  may not result in competitive exclusion
• However, numerous studies have not supported this
  hypothesis originally

16
Size RatiosEvolutionary Extinction Morphology

• One problem with analyzing present-day
  assemblages is extinctions or shifts in
  morphology are not observed but must be inferred
  indirectly

17
Size RatiosEvolutionary Extinction Morphology

• Fossil assemblages can give insight into
  long-term patterns of extinction and morphology
• For example, are extinctions random with respect
  to morphology or are species that are too close
  to one another more likely to go extinct?

18
Size RatiosEcological Extinction Morphology

• In ecological time, can extinctions be predicted
  on the basis of morphology or other species
  attributes?
• Many (e.g. conservationists) would be very
  interested in knowing if extinctions are random
  with respect to body size, habitat affinity, or
  trophic status

19
Size RatiosExample of Ecological Extinction

• Moulton and Pimm argued that extinction of
  introduced species in the Hawaiian Islands could
  be predicted on the basis of body size or
  morphology, chiefly resulting from competition
  from introduced species

20
Size RatiosExample of Ecological Extinction

• For example, introduced pairs of congeneric
  species that both survived on at least one island
  (n6) differed more in bill length (22) than
  congeneric pairs in which one of the pair went
  extinct (9 n9)
• Introduced species were overdispersed in
  morphological space compared to random draws of
  species from the set of all introduced forest
  passerines

21
Size RatiosExample of Ecological Extinction

• Introduced finches of Oahu and introduced
  passerines in Tahiti also exhibited morphological
  overdispersion
• Fig 6.12. Surviving species
• (filled squares) were significantly
• overdispersed in comparison
• to random communities
• However, not so
• clear-cut

22
Size RatiosEmpirical Tests

• The size ratio controversies have not only
  addressed size overlap and better statistical
  tests, but they have begun to account for other
  sources of variation known to influence the
  mechanisms driving these patterns (e.g.
  geographic variation in morphology, patterns of
  resource use)

23
Size Ratios Empirical Tests

• Ecological Character Displacement in the Red Fox
  (Dayan et al. 1989)
• Competition and Morphology of Co-occurring
  Dytiscid Beetles (Juliano and Lawton 1990)
• Bill Sizes of Galapagos Finches (Lack 1947, Grant
  1972)

24
(No Transcript)