of species and used a gradually increasing rate function to arrive at the extinction estimates in the Global 2000 Report (Table 6–2). Ehrlich and Ehrlich (1981, p. 280) assumed that “the diversity of species will be lost more rapidly than the forest itself and used an exponential function to estimate depletion of species. They assigned a constant rate of increase to the rate of depletion based on human population growth (1.5% per year), human impacts in overdeveloped countries (1% per year), and forest loss (1% per year). The total rate of increase (3.5% per year) plus an assumed current rate of species extinction (1% per year in one calculation and 2% in another) were substituted in the exponential function to obtain the estimate of species depletion (Table 6–2).

The rates of deforestation used in both estimates discussed above are 3.8 to 5.5 times higher than the rates obtained by Lanly (1982). If Lanly’s values are substituted in Lovejoy’s analysis (Table 6–3), the estimate of species extinctions by the year 2000 would be almost 9% of the total biota instead of 33 to 50%. The high estimate of Ehrlich and Ehrlich would be halved simply by changing the assumed fraction of the biota presently undergoing depletion. The function used by the Ehrlichs is very sensitive to changes in assumptions because of its exponential nature and the absence of any negative feedback to stabilize its response. Therefore, any change in the value of any of the factors contributing to the rate of increase or the rate of extinctions would change the prediction significantly (Table 6–3). For example, if the estimated rate of avian extinctions in Puerto Rico (discussed below) is substituted for the total rate of extinctions, the expected species depletion by the year 2010 would be reduced to 4%. The estimates by the Ehrlichs also suffer from not taking into account the heterogeneity of destruction and regeneration among different forest types. Are these definitive estimates? Clearly not!

Correcting for differences in species richness of forests, forest recovery rates, and differential human impact by forest type will certainly lower any of the estimates that now lack consideration of mitigating factors. Furthermore, the functions used to relate forest loss to species loss are still to be established experimentally. When and if this comes about, the results may be either more or less conservative than those assumed by either Lovejoy or Ehrlich and Ehrlich.

Lower extinction rates for plants (Table 6–1) were estimated by Simberloff (1986) by using a species-area relationship, conservative assumptions about the fraction of forest area loss, a Z factor (an exponent of the forest area lost) of 0.25, and various scenarios of forest conservation. Simberloff could not derive a mass extinction of plant species by the year 2000 comparable to those of the geological past, even though his analysis does not correct for forest recovery after conversion. However, his estimates of extinction are lower than those discussed above, even though the function he used usually accounts for only 44.8% of the variation in species when area changes.


I believe that to estimate the reduction in the number of species in the tropics it is necessary to consider the effect of forest types on species abundance, the spatially selective (life zone) intensity of human activity, the role of secondary

The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement