Limitations of these analyses

• Projections vs. forecasts
  • A projection says what would happen if current trends continue.

  • A forecast says what will happen.

  • It is best to regard these analyses as projections, because we know that populations can't grow at a geometric rate forever. They are tools for assessing the current condition of the population and for identifying critical stages.

• Large statistical uncertainties associated with estimates of leading eigenvalue.

• Elasticities vs. sensitivities vs. management possibilities
  • Elasticities provide a sense of proportional effect.

  • Sensitivities provide a sense of numerical effect.

  • What we really need is a measure of effect over the range of possible management options.

• Frequent focus on leading eigenvalue may be misplaced. It indicates the asymptotic properties of population growth rate, but the trend in the next few generations may be much more important (and much different).

• No environmental or demographic stochasticity

• Single isolated populations

Still, these analyses are useful.

• Only under very special conditions can a species be viable long-term unless at least one of its constituent populations is also viable.

• Help identify focus for management efforts. If individual populations are viable, focus should be on protection of habitat. If individual populations are not viable, some of them must also be managed to restore viability (or at least to prevent extinction).

• Provide a framework to which extensions including environmental and demographic stochasticity and migration among populations is easily added.

• Provide a focus for identifying critical life-history stages, even in the absence of a complete demographic analysis. Identify the life-history stages at which intervention is most likely to be useful.