M.S. Thesis

		Shefferson, R.P. 2000. Demographic applications of mark-recapture models to
		populations of rare herbaceous perennials. M.S. thesis, University of California at Berkeley, Berkeley, California. 95p.

Abstract Demographic research has become integral to conservation and evolutionary
biology (Harper and White 1974, Charlesworth 1980, Schemske et al. 1994). Robust
estimates of survival, fecundity, and fitness are critical in such diverse situations as
monitoring of rare species (Beissinger and Westphal 1998), assessment of invasion
potential and impacts for weedy species (Marschall and Crowder 1996, Parker and
Reichard 1998), and investigation of life history trade-offs (Stearns 1992). Yet features
of many plants, including some rare species, can present serious obstacles to
population-oriented demographic studies.

A phenomenon that is particularly challenging to studies in plant demography is
dormancy. Dormancy is a 'state of suppressed development' (Danks 1987) that allows
an organism to conserve energy and survive harsh periods (Dahms 1995), and occurs
in many taxa (Lyman 1982, Wang 1989, Geiser 1994, Dahms 1995). It appears to be a
response to changing environmental factors. Dormancy has been studied extensively
in animals but less often in plants. In general, it has been considered a bet-hedging
strategy (Seger and Brockman 1987, Philippi and Seeger 1989). However, the
proximate causes of dormancy remain poorly understood. Particularly so in geophytes,
a group of herbaceous plants defined by the presence of a perennating subterranean
storage organ (Lesica and Steele 1994). In these plants, adult dormancy is a novel
system to examine potentially complicated life history trade-offs involving survival,
sprouting, and reproduction.

Underground dormant stages in mature genets of geophyte species have received little
attention in plant demography, but have been noted recently as sources of bias (Gilbert
and Lee 1980, Lesica and Shelly 1995, Rasmussen and Whigham 1998). Here, adult
dormancy becomes a phenomenon mimicking death. The bias from dormancy lies in
the assumption that all live plants can be detected, and can affect population counts,
survival estimates, and other demographic parameters (Gilbert and Lee 1980, Menges
1986). While little bias may result if only a consistently small percentage of the
population exhibits an underground life stage, the potential for bias increases with
increasing ubiquity of the phenomenon. For endangered plants, which are likely to
occur in small populations, the potential for bias may be high due to random chance.
The likelihood of resprouting, which is a measure of the tendency of an individual to
sprout in the growing season following the last sprouting event, has been used as an
estimator of survival. However, this measure is not robust to this source of fluctuation in
observable population size.

Terrestrial orchids, such as the small yellow lady's slipper (Cypripedium calceolus ssp.
parviflorum (Salisb.) Fernald), exemplify the major problems inherent in demographic
studies of rare geophytes. First, small population size and rarity can limit sample size
and hence power, resulting in high uncertainty in demographic estimates (Travis and
Sutter 1986). This uncertainty can render population viability analyses too poor to be of
value in management (Beissinger and Westphal 1998). Second, rarity can also make
destructive experimental methods impractical and unethical (Travis and Sutter 1986).
Third, seed dormancy and miniscule to microscopic seed sizes can present serious
challenges to the estimation of fecundity and fitness, making it difficult to test life history
hypotheses. Lastly, underground dormant stages may prevent the accurate censusing
of populations, and may also bias survival estimates (Rasmussen and Whigham
1998).

The shortcomings of conventional survival estimation may be addressed with a
versatile family of statistical techniques known as mark-recapture models. Alexander et
al. (1997) adapted the closed population modeling strategy of Otis et al. (1978) to
develop estimators of population size unbiased by dormancy. Here, I present novel
extensions of this theory to the estimation of survival, dormancy, and life-stage
transitions in the small yellow lady's slipper. Furthermore, I utilize these techniques to
answer basic life-history questions involving costs of reproduction and sprouting.

The techniques proposed in this thesis are versatile and have already been applied to
fields as diverse as animal population biology (Lebreton et al. 1992), metapopulation
biology (Erwin et al. 1998), and paleobiology (Nichols and Pollock 1983, Rosenzweig
and Clark 1994). In the now classic paper of Harper and White (1974), the authors
suggest that since plants and animals are morphologically distinct from one another in
demographically important ways, the dynamics of their populations must also be
distinct. They further suggested that plant demography and animal demography occupy
different theoretical and methodological spheres: "A science of plant demography
cannot be built uncritically in the methods developed by zoologists" (Harper and White
1974: 420). Although the theory of plant demography necessarily involves conceptual
challenges fundamentally different from animal demography, an interdisciplinary
approach between the two fields can still yield remarkable rewards. In this thesis, I
explore the applicability to plants of statistical approaches used to estimate annual
demography. Hopefully, the approaches presented here will yield advances to
computational demography, conservation biology, and life history theory.

Thesis copyright notice: © 2000 by Richard P. Shefferson

M.S. Thesis