  Understanding Rarity
© Eric R. Pianka 
Understanding rarity constitutes a major challenge for ecologists.
Main (1982)
asked "Are rare species precious or dross? And, are they vital to community function?"
Do rare species persist in more stable communities in spite of their rareness, or does
the presence of rare species enhance the stability of ecosystems? Main (1982) suggested
that one reason so many rare species exist may be that ecosystems have been "over-written
many times after imperfect erasures" (incomplete extinctions). Consequently, current
ecosystems contain numerous relicts of their predecessors assembled under different conditions.
Main suggested that rare species could be vital to long-term ecosystem sustainability, providing
'insurance' for the delivery of ecosystem functions by alternative means in the event of drastic
environmental changes.
Most species of Australian desert lizards are uncommon, making them difficult to study.
Some are extremely rare to the point of vanishing rareness. Regardless of how rareness is
defined, most ecologists concur that the majority of species are indeed uncommon (Gaston 1994,
Kunin and Gaston 1997). Magurran and Henderson (2003) distinguished between relatively abundant
"core species" and uncommon "occasional species". Chronic rarity has proven to be exceedingly difficult
to study, but, as mentioned above, rare species could well be very important to community
function (Main 1982, Morton and
James 1988; Kunin and Gaston 1997; Preston 1948, 1962; Thompson et al. 2003).
Over the last 42 years, I have witnessed metapopulation-like local extinctions and colonization in a few species. I have captured migrants of a number of species (Ctenophorus fordi, Ctenophorus scutulatus, Ctenotus greeri, Ctenotus leae, Ctenotus leonhardii, Lophognathus longirostris, Nephrurus vertebralis, and Tiliqua occipitalis) dispersing through habitats that they do not normally occupy. Until recently, inadequate sample sizes have prevented me from doing much with uncommon species, but I have now finally managed to acquire large enough samples to attempt to understand the ecologies of most of them -- no one else has ever managed to collect such large samples of uncommon Australian desert lizards. 
Several hypotheses for the continued existence of such rare species spring to mind: H 1 (the niche breadth hypothesis). Rare species are uncommon because they are specialized with narrow dietary or microhabitat niche requirements. Resources such as habitats, microhabitats, or foods might be scarce or limited. These alternatives can be tested with data on niche breadths (below). H 2. Rare species could have narrow geographic ranges, occurring at only a few sites. H 3. Rare species might have narrow tolerances to physical environments. H 4. Rare species might be uncommon only locally in 'sink' populations, but might be more abundant in nearby 'source' areas. H 5. Rare species could be rare because they do not have dispersal powers necessary to find and invade suitable habitats. H 6. Predators could hold population densities of uncommon species at low levels. H 7. Rare species could be uncommon due to diffuse competition from many other, more abundant, species (MacArthur 1972b). Some related questions that can be asked about rare species include: How can rare species find mates and continue to exist?
Could some species be uncommon because of their low fecundity?
Are rare species merely accidentals, dispersing from one habitat to another?
How important are rare species to the function and stability of communities?
Is rarity an illusion due to cryptic behavior making putative rare species difficult to find?
Merely being in an alien habitat is not necessarily a death sentence, as these habitats offer shelter and food -- a migrant that succeeds in reaching its correct habitat could also reap the benefits of sweepstakes reproductive success. Which of these factor(s) is/are crucial determinants of commonness or rarity needs to be determined for each species. 
Abundances of 15 species of Ctenotus skinks plotted against the size of their geographic ranges. Widespread species are not necessarily common, but species with narrower ranges such as Ctenotus quattuordecimlineatus can be much more abundant locally. Hence, H 2 (above) is not supported. 
Histograms showing the number of abundant species (blue), species of intermediate abundance (green), and rare species (red) found at 10 different study sites in the Great Victoria Desert. Rare species tend to be found on fewer sites, but one rare species is found on all 10 study areas. Species of intermediate abundance are distributed bimodally, with some occurring at only a few sites but others are present on most sites. Abundant species are also bimodally distributed and occur at a greater number of sites on average. These data also do not support H 2 (above). 
Australian desert lizards occur at four different habitat types in sandridge deserts: crest, slope, base, and flat (shown in above figure). Some species are largely restricted to one of these habitats, but other species display more generalized habitat requirements. Habitat niche breadths were computed for each species based on proportional representation in each of the four habitats using Simpson's index of diversity, 1/Σpi2, where pi is the proportion of lizards found in habitat i. Habitat niche breadths range from 1 to 4 (below). 
Histograms of habitat niche breadths of abundant species (blue), species of intermediate abundance (green), and rare species (red). Many rare species are habitat specialists, as are some intermediate and a few abundant species. Habitat niche breadths are broadest in abundant species and those of intermediate abundance, although 6 rare species also have broad habitat niches. In all three abundance categories, species exhibit the full range of habitat niche breadths. The niche breadth hypothesis (H 1 above) is not strongly supported by these data on habitat niche breadths. 
Each lizard was assigned to one of 15 different microhabitat resource categories. Microhabitat niche breadths were computed using Simpson's index of diversity. The above figure shows histograms of microhabitat niche breadths for abundant species (blue), species of intermediate abundance (green), and rare species (red). Most abundant species and many rare species are microhabitat specialists, as are a few species of intermediate abundance. Microhabitat niche breadths are narrowest in abundant species. Many species of intermediate abundance exhibit broad microhabitat niches. A few rare species are actually microhabitat generalists. Hence, again, the niche breadth hypothesis (H 1 above) seems to be refuted by these data on microhabitat niche breadths. 
To estimate dietary niche breadth, 20 prey categories were recognized, corresponding mostly to arthropod orders plus categories for plant and vertebrate food items. Simpson's index was again used to estimate dietary niche breadths. Species in all three abundance categories exhibit a wide range of food niche breadths. As with microhabitat niche breadths, dietary niche breadths on average are narrowest in abundant species and broadest in species of intermediate abundance. Abundant species and those of intermediate abundance have bimodal distributions, with some relatively specialized species and others with broader diets. Rare species are distributed more uniformly, with a complete range of food niche breadths. Hence, once again, the niche breadth hypothesis H 1 is not supported by these data on dietary niche breadths. 
Microhabitat niche breadths are weakly positively correlated with dietary niche breadths in all three abundance categories, although a great deal of scatter exists. In all three abundance categories, some species are both dietary and microhabitat specialists. Several species of intermediate abundance exhibit narrow diets but relatively broad microhabitat niche breadths, whereas others, including several rare species, have broad niches on both dimensions. 
To attempt to test the diffuse competition hypothesis (H7), the logarithm of abundance is plotted against the average total dietary overlap with all other species in the above figure. The correlation goes against the prediction with higher overlaps among more abundant species. In the next table, abundances and values of each of nine variates are listed for each of the 12 most abundant species, along with averages for all abundant species plus those for all species. Values that deviate above or below expected values are highlighted in bold. As expected, all the abundant species are small and all but a couple are widespread, occurring on most sites. Also, as expected, most tend to have broad niches though some are specialists. 
In the following table, abundances and values of each of nine variates are listed for each of the 18 species of intermediate abundance, along with averages for all intermediate species plus those for all species. Again, values that deviate from expected are highlighted in bold. 
The following table lists abundances and values of each of nine variates for each of 33 rare species, along with averages for all species. Values that deviate as expected for rare species are highlighted in bold. Eight rare species are large. Most, but not all, are found at few sites. Eleven experience high dietary overlap, hence presumed diffuse competition. Many rare species have narrow niche breadths. Each and every rare species displays at least one value expected to be associated with rarity. Many exhibit several values as predicted. 
The following table summarizes correlations and partial correlations among variates with the logarithm of abundance. The three strongest correlations (shown in bold) are with the number of sites, dietary overlap, and habitat niche breadth. Results of stepwise multiple regression with the logarithm of abundance as the dependent variable are given below the table. These 3 variables reduce the variance in the logarithm of abundance by 59.8%. 
The first two principal components (shown below) of a PCA analysis based on the same 9 variates capture 46.3% of the variance. Rare species tend to be separated from common species and those of intermediate abundance. Adding PC3 and PC4 (not shown) reduces residual variance by a further 26.9%, for a total of 73.2%. 
A discriminant function analysis based on the same 9 variables clearly separates the 12 abundant species from the less common species. Some overlap occurs between rare species and those of intermediate abundance (figure below), but species in these two abundance categories display some separation. 
While some species may be rare because of narrow niche requirements, other specialized species can be abundant. No general explanation for rarity may exist, but rather each species may have its own idiosyncratic reasons for being uncommon. Many of the possible factors that could contribute to rarity remain to be evaluated, and the difficulty of studying uncommon species remains a formidable challenge to ecologists. Caveat: Foregoing analyses are overly simplistic in that they assume abundances and niche breadths remain fixed in time, but of course, they must vary. These lizard populations can be viewed more realistically using a metaphor: their relative abundances are analogous to 3-dimensional waves in as many dimensions as there are species: the vertical coordinate represents the abundances of each of the various species moving up and down and around in space and time. At any given time, some species are abundant while others may be scarce. These abundance surfaces are relatively placid for some species, but very rough for others. Some species like Ctenophorus clayi and Ctenophorus nuchalis boom and bust, exhibiting intermittent rarity, whereas others like Cyclodomorphus, Eremiascincus, and Tiliqua, are always uncommon (chronic rarity). Still other common species, like Ctenophorus isolepis, exhibit more stable populations. These waves of relative abundance respond to fire and episodic precipitation events (see Fire Succession on the B-area), both of which drive changes in resource availabilities of prey and microhabitats through time and space. One of my long-term goals is to attempt to describe this multidimensional spatial-temporal wave-like landscape. References Gaston, K. J. 1994. Rarity. Population and Community Biology Series 13. Chapman and Hall. Kunin, W. E. and Gaston, K. J., eds 1997. The biology of rarity: causes and consequences of rare-common differences. Chapman and Hall. MacArthur, R. H. 1969. Species packing and what interspecific competition minimizes. Proc. Nat. Acad. Sci. USA 64: 1369-1371. MacArthur, R. H. 1970. Species packing and competitive equilibrium for many species. Theoret. Population Biol. 1: 1-11. MacArthur, R. H. 1972a. Strong, or weak, interactions? Trans. Conn. Acad. Arts and Sci. 44: 177-188. MacArthur, R. H. 1972b. Geographical Ecology. 269pp. Harper and Row. Magurran, A. E. and P. A. Henderson 2003. Explaining the excess of rare species in natural species abundance distributions. Nature 422 (6933): 714-716. Download pdf. Main, A. R. 1982. Rare species: precious or dross? Graves, R. H. and W. D. L. Ride (eds): Species at risk: Research in Australia, pp. 163-174. Australian Academy of Science, Canberra. Download pdf Morton, S. R., and C. D. James. 1988. The diversity and abundance of lizards in arid Australia: a new hypothesis. American Naturalist. 132: 237-256. Download pdf. Pianka, E. R. 2011. Notes on the ecology of some uncommon skinks in the Great Victoria Desert. Western Australian Naturalist 28: 50-60. Download pdf Preston, F. W. 1948. The commoness and rarity of species. Ecology 29: 254-283. Preston, F. W. 1962. The canonical distribution of commoness and rarity. Ecology 43: pp. 185-215 and 410-432. Simpson, E. H. 1949. Measurement of diversity. Nature 163: 688. Thompson, G. G., P. C. Withers, E. R. Pianka, and S. A. Thompson. 2003. Assessing biodiversity with species accumulation curves: Inventories of small reptiles by pit-trapping in Western Australia. Austral Ecology 28: 361-383. free website hit counter code |