Catkin

Abstract: Mountain birch is a subarctic tree that faces a high risk of severe attack from herbivores. During outbreaks of its major pests, geometrid moths, a typical sequence of defoliation lasts 2-3 successive years. We simulated a defoliator outbreak by different combinations of 25% and 75% defoliations over three successive years. We studied patterns of resource allocation during and after the simulated outbreaks by quantifying the effect of different degrees of cumulative defoliation on mountain birch resistance, growth, and reproduction. Bioassays showed that the delayed inducible resistance largely waned within 2 yr, whereas the effects of defoliation on tree growth and reproduction persisted longer. Cu- mulative defoliations intensified birch resistance only slightly, and the efficacy of resistance remained in the range induced by a single defoliation event in other studies. This suggests that, even after cumulative defoliations, the "killing power" of induced resistance in moun- tain birch accounts for only a fraction of that required to end a geometrid outbreak. A low level of defoliation (25%) did not generally affect the trees, but high (75%) defoliation decreased their growth and reproduction and also decreased their palatability to insects. Sequential annual defoliations reduced the mass of female catkins, whereas seed- specific mass and seed viability were most affected by previous-year defoliation. The num- ber of catkins per tree was most affected by defoliation 2 yr earlier. Thus, previous-year defoliation affected catkin and seed properties due to the exhaustion of nutrients required for their development, but earlier defoliation limited the catkin number by affecting bud type determination. Similarly, long shoot number was affected by defoliation 2 yr earlier, but cumulative defoliations also reduced the length of long shoots. Trees maintained leaf biomass at the expense of growth and reproduction; leaf dry mass, leaf area, and specific leaf mass were only slightly and nonsignificantly reduced by previous-year defoliation. For mountain birch, maintenance of normal leaf biomass may be a tolerance strategy for coping with herbivory that is at least as important to a long-lived plant as is induced resistance.

Abstract: We studied the relative roles of insects and wind as pollen vectors for the arroyo willow, Salix lasiolepis Benth. through insect exclosure studies at two elevational sites. Insect pollination resulted in production of 99.9% of seeds at a low elevation site (2,130 m) and at a high elevation site (2,450 m) with wind playing an exceedingly small role. Few pollen grains reached individual pistillate catkins despite the presence of abundant airborne willow pollen near male plants. In flowers that showed evidence of receipt of wind transferred pollen, due to their swollen appearance, we found only 1.74 fertilized ovules per flower and that 90.9% of these fertilized ovules were aborted by the maternal plant. In contrast, flowers open to insect visitation possessed 7.00 fertilized ovules per flower, 83.3% of which became filled seeds. We tested for the existence of apomictic seed production through use of closed bags that excluded both insect-and windtransferred pollen but found no evidence of seed produced in this treatment. Additional control studies documented that wind treatment bags did not impede pollen flow while neither wind or closed bags influenced seed set in our treatments. THE interrelation between floral morphology and pollination syndrome has been central to arguments regarding angiosperm evolution (Leppik, 1957; Pijl, 1960; Meeuse, 1972; Baker and Hurd, 1968; Regal, 1977). The primacy of entomophily, or insect pollination, in the success of angiosperms is accepted in most contemporary systematic schemes (Cronquist, 1981; Crepet, 1984) although alternatives supporting anemophily as primary in some early angiosperms are still presented (Meeuse, 1972). The family Salicaceae has been important in debates over the nature of the primitive pollination mode in angiosperms because at different times the family has been considered either primitive or advanced based on the simI Received for publication 16 July 1987; revision accepted 6 January 1988. Research was supported by NSF grants DEB-80221754, BSR-8314594, and DEB-8205904, an NAU Organized Research grant, and grants from Sigma Xi and the ARCS Foundation to CFS. We thank P. Thompson, Director, and A. Phillips, Curator of Biology, for permission to conduct research on Museum of Northern Arizona property. We are grateful for taxonomic identifications of Andrenidae by W. LaBerge of the Illinois Natural History Survey and of Halictidae by G. Eickwort of Cornell University. R. Hevly discussed willow pollen sampling with us at length. G. Bell provided valuable statistical advice. W. L. Montgomery, T. G. Witham, A. White, T. P. Craig, and R. S. Fritz provided helpful comments on ideas in this study. This research was completed in partial fulfillment of requirements for the Ph.D. degree at NAU by CFS. 2 Current address: Blandy Experimental Farm and Dept. of Environmental Sciences, Clark Hall, University of Virginia, Charlottesville, VA 22903. plified nature of its catkins (Fisher, 1928; Eames, 1961). Eames (1961) suggested that the simplicity of the Salicaceae is due to reduction. Currently, the family Salicaceae is considered moderately advanced, with the genus Populus producing anemophilous catkins considered derived from entomophilous ancestors (Fisher, 1928; Dorn, 1976) and the more advanced genus Salix producing catkins that have secondarily reacquired entomophily (Stebbins, 1970; Cronquist, 1981). The pollination mode of the dioecious Salix has been considered transitional between entomophily and anemophily. Salix produces apetalous flowers combined into catkins reminiscent ofanemophilous inflorescences but also possesses traits typical of entomophilous flowers such as nectaries and sculpted pollen (Wodehouse, 1935; Percival, 1965; Argus, 1974). Most authors have accepted that willows are primarily entomophilous but frequently have suggested a facultative role for wind pollination (Eames, 1961; Faegri and van der Pijl, 1971; Proctor and Yeo, 1973). The prediction of wind pollination was a result of the abundance of willow pollen in atmospheric pollen samples (Proctor and Yeo, 1973; Lewis, Vinay, and Zenger, 1983) in addition to simplified catkin structure. Experiments to elucidate the contribution of insects or wind to pollination in species that appear to straddle the borderline between insect and wind pollination must be performed on a species-by-species basis (Proctor, 1978). The need for species-specific analysis of pol-

Abstract: Most damage caused by invertebrate herbivores is local and limited in extent and may therefore seem inconsequential. However, in addition to providing photosynthates to the whole plant, individual leaves may preferentially feed local sinks, including pri- mordial meristems that develop later. For leaf damage, this may have local consequences and, in the case of preformed meristems, may result in time lags. To investigate such consequences and to determine the degree of independence among shoot modules, we studied shoot traits after damage to individual leaves in long shoots and vegetative and generative short shoots of field-growing mountain birch, Betula pubescens subsp. czere- panovii (Orlova) Hamet-Ahti. For short shoots, the results suggest that, due to the catkin, generative shoots have a higher priority in resource use than vegetative shoots. The overwinter survival of vegetative short shoots was decreased after their young leaves had been damaged early in the previous season. Otherwise, there were no significant effects of leaf damage on vegetative shoots. In generative shoots, leaf damage affected size and reproduction instead of survival. Because leaf damage significantly decreased leaf size-in the generative shoots in the posttreatment year, leaves presumably competed with the catkin for resources within the shoot. Damage also reduced the catkin mass and the mean seed mass in the treatment year. In the post- treatment year, catkin mass and total seed mass were reduced by the treatments, but dif- ferences in the number of seeds and mean seed mass did not quite reach significance. The treatments had no effects on leaf mass of neighbors of vegetative or generative shoots in the treatment year or in the posttreatment year. In the treatment year, removal of long shoot leaves from a growing long shoot did not affect growth of the long shoot, indicating that such growth was supported by resources external to the shoot. However, local effects were evident in the posttreatment year: leaf mass decreased in shoots that developed from axillary buds in long shoots whose supporting leaves had been clipped in the previous year. Consequently, removal of individual leaves may have local effects on shoot survivorship, reproduction, and growth. Timing of damage and shoot type modified the consequences of local leaf damage, and there were no consistent responses for all situations. Due to the limited extent of damage, the effects were not likely to result from nutrient loss due to damage. Instead, consequences of local leaf damage could be explained largely by effects on local sinks and meristem primordia.