J. Nunezfarfan et R. Dirzo, EVOLUTIONARY ECOLOGY OF DATURA-STRAMONIUM L IN CENTRAL MEXICO - NATURAL-SELECTION FOR RESISTANCE TO HERBIVOROUS INSECTS, Evolution, 48(2), 1994, pp. 423-436
It has been assumed that herbivores constitute a selective agent for t
he evolution of plant resistance. However, few studies have tested thi
s hypothesis. In this study, we look at the annual weed Datura stramon
ium for evidence of current natural selection for resistance to herbiv
orous insects. Paternal half-sib families obtained through controlled
crosses were exposed to herbivores under natural conditions. The plant
s were damaged by two folivorous insects: the tobacco flea beetle Epit
rix parvula and the grasshopper Sphenarium purpurascens. Selection was
estimated using a multiple-regression analysis of plant size and of d
amage by the two herbivores on plant fitness measured as fruit product
ion for both individual phenotypes and family breeding values (genetic
analysis). Directional phenotypic selection was detected for both lar
ger plant size and lower resistance to the flea beetles, whereas stabi
lizing phenotypic selection was revealed for resistance to S. purpuras
cens. However, performing the same analyses on the breeding values of
the characters revealed directional and stabilizing selection only for
plant size. Thus, no agreement existed between the results of the two
types of analyses, nor was there any detectable potential for genetic
change in the studied population because of selection on herbivore re
sistance. The narrow-sense heritability of every trait studied was sma
ll (all < 0.1) and not different from zero. The potential for evolutio
nary response to natural selection for higher resistance to herbivores
in the studied population of D. stramonium is probably limited by lac
k of genetic variation. Natural selection acts on phenotypes, and the
detection of phenotypic selection on resistance to herbivores confirms
their ecological importance in determining plant fitness. However, ev
olutionary inferences based solely on phenotypic selection analyses mu
st be interpreted with caution.