We used simple linear regressions to determine how soil moisture affected total pollen deposition, total pigment deposition, pollen deposition per open male flower, and pigment deposition per open male flower on bee-pollinated stigmas. Pollen deposition per open male flower was right-skewed and thus inverse transformed to improve normality. Pigment deposition and pigment deposition per open male flower from high-irrigation plants were right-skewed and thus square-root transformed to improve normality. Pigment deposition and pigment deposition per open male flower from low-irrigation plants were right-skewed and thus log10 transformed to improve normality of the residuals. We used a multiple logistic regression to assess how soil moisture and pollination affected fruit set, and a general linear model to assess how soil moisture and pollination affected seed set. To assess the effect of bee visitation on fruit set, we used a multiple logistic regression containing two fixed factors: Apis mellifera visitation rate and Eucera visitation rate. For this analysis, plastic pots large we used visitation rates from surveys of bees visiting all plants in the plot .
To determine the effect of pollen and pigment deposition on seed set in bee and hand-pollinated plants, we used separate general linear models containing total pollen deposition or proportion pigment from high-irrigation plants as the fixed factor.The irrigation treatment affected Cucurbita pepo plants to some degree in terms of floral traits. Male flower production and pollen production linearly increased with increasing soil moisture , but no other measured floral trait exhibited a significant response to manipulated soil moisture levels . Video recordings confirmed that honey bees and squash bees were the most frequent visitors of squash flowers . Although there were no significant effects of plant soil moisture on the visitation rate of Apis and Eucera on male flowers , we did observe contrasting patterns of visitation on female flowers . Summed across Apis visits, both the time spent drinking nectar and contacting floral stigmas increased with plant soil moisture, as well . For Eucera, in contrast, visitation rate on female flowers was independent of soil moisture ; cumulative time spent drinking nectar and contacting floral stigmas followed suit . Pollen and pigment deposition on the stigmas of bee-pollinated plants increased with increasing plant soil moisture .
For pollen deposition we observed a linear increase with increasing plant soil moisture , but this relationship was no longer significant when we controlled for the number of open male flowers available to bees on the day that each stigma was collected . Pigment originating from male flowers on plants in the high-irrigation group was deposited in increasingly greater amounts on the stigmas of female flowers as the soil moisture experienced by recipient plants increased . This relationship remained significant after we controlled for the availability of male flowers from high-irrigation plants on the day that each measurement was made . For fluorescent pigment originating from male flowers on plants in the low-irrigation group, pigment transfer to stigmas was independent of the recipient plant’s soil moisture . Pollen and pigment deposition for hand-pollinated stigmas are listed in Appendix 2.2. Fruit set was independent of plant soil moisture and also did not differ as a function of pollination mode : bee-pollinated mean and hand-pollinated mean . For bee-pollinated plants, however, the likelihood of fruit set increased with increasing honey bee visitation but not with increasing squash bee visitation . Seed set increased with plant soil moisture and was higher for bee-pollinated plants than for hand-pollinated plants . For bee-pollinated plants, we further observed that seed set increased with the amount of pollen deposited on stigmas and with the proportion of pigment originating from plants in the high-irrigation group . In contrast, neither relationship was significant for hand-pollinated plants .We tested three predictions concerning how specialist and generalist pollinators interact with plants grown under a gradient of soil moisture conditions and how pollinator-specific behaviors affect plant reproduction.
We received partial support for our first prediction. The frequency of floral visits by honey bees increased with increasing plant soil moisture for female flowers but not for male flowers; visit frequency by squash bees was independent of plant soil moisture for both female and male flowers. Moreover, increased visitation by Apis in female flowers included increased duration of stigmatic contact with increasing plant soil moisture. Our second prediction was met; stigmatic deposition of pigment from plants in the high-irrigation group increased with increasing plant soil moisture. Although we were not able to directly track pollen deposition from plants in the high-irrigation group, pigment and pollen deposition were highly correlated. Support for our third prediction comes from the comparisons of bee-pollinated versus hand-pollinated plants. As expected, seed set increased with increasing plant soil moisture for both bee-pollinated and hand-pollinated plants. For bee-pollinated plants, seed set increased with increasing pollen deposition and with increasing pigment from plants in the high-irrigation group. Hand-pollinated plants, in contrast, received an even mix of pollen from plants in the low-irrigation and high-irrigation groups irrespective of the plant soil moisture of the recipient plant . These plants did not exhibit relationships between seed set and pollen deposition nor between seed set and the amount of pigment originating from plants in the high-irrigation group. These results taken together suggest that pollen from plants in the high-irrigation group was of higher quality than pollen from plants in the low-irrigation group .Given that the goals of this study were to investigate how the behavior of pollinators mediates the reproductive performance of plants that experience a gradient of soil-moisture environments, we did not attempt to identify the mechanisms responsible for why pollinators exhibit selectivity. Although an earlier study in this same system found that flower size and nectar production increased with increasing plant soil moisture, we found no evidence for similar responses in the current study and assume that inter-annual differences in temperature , the timing of planting, or differences in the experiment itself are responsible. One possible explanation for our finding of increased visitation by honey bees to female flowers with increasing plant soil moisture may have been caused by the emission or composition of floral VOCs attractive to Apis changing with increasing soil moisture . Irrespective of the mechanism of attraction, Apis often exhibit a preference for female flowers of C. pepo as well as those of C. foetidissima . Given that pollen production increased in male flowers with increasing soil moisture, increased visitation to female flowers with increasing soil moisture by honey bees is sufficient to explain higher deposition of pollen from plants in the high-irrigation group given that exhibited cumulative duration of stigmatic contact increased with visitation. We relied on fluorescent pigments to examine pollen movement among plants in this experiment. As a proxy for pollen deposition, black plastic nursery pots the use of pigments is established in the literature and in our study seems well justified given the close correlation between pollen and pigment deposition .
Nonetheless, this approach, by itself, does not permit direct identification of pollen from plants in low-irrigation and high-irrigation groups once it is deposited on the stigma. In fact, total stigmatic pollen deposition was independent of plant soil moisture after correcting for male flower availability. This result differed from the positive effect of increased soil moisture of the recipient plant on the deposition of pigment from plants in the high-irrigation group after correcting for available male flowers. However, total pollen deposition included pollen from plants in both the low-irrigation and high-irrigation groups. Therefore, the inclusion of pollen from all soil moisture levels in the measurement of total pollen deposition likely canceled out the effects from the high-irrigation pollen, as we found that pigment from plants in the low-irrigation group did not increase with increasing plant soil moisture of the recipient plant . As with our analysis of pollen deposition, analysis of fruit set also did not allow for an assessment of the contribution of pollen from different sources. It is noteworthy, however, that fruit set increased with increasing visitation by honey bees and was independent of visitation by squash bees. Other work in this system has revealed that Eucera outperform Apis with respect to per visit contributions to fruit set . Thus, the contributions of honey bees to fruit set in the present study provides additional evidence that Apis transport pollen from male flowers of plants in the high-irrigation group to the female flowers of plants in the high-irrigation group and in doing so enhance plant reproduction. Squash bees, in contrast, indiscriminately visited flowers with respect to plant soil moisture and likely transported pollen from plants across the soil moisture gradient. The findings of this study seem likely to apply to other plant-pollinator systems. Generalist bees often forage preferentially for high-quality floral resources, such as more viable pollen , pollen with a high protein content , nectar of increased volume and concentration , and nectar with increased amounts of amino acids . These preferences, in turn, can benefit plants through increased pollinator visitation and pollen deposition. Kasagi and Kudo , for example, found that preferences exhibited by bumble bees for increased nectar resources resulted in reduced pollen limitation in Phyllodoce. However, generalists need to not only be attracted to high-quality resources, but they also need to increase stigmatic deposition of high-quality pollen to augment plant reproduction. Mu et al. , for example, assessed that preferences for flower color by flies and honey bees coincided with increased pollen viability, which led to greater reproductive success in Gentiana leucomelaena. Compared to generalists, specialist pollinators are often considered more effective . Yet during stressful environmental conditions, such as droughts, generalists may preferentially visit high quality resources while specialists do not. Exhibiting distinct floral constancy amongst conspecifics of high-quality could thereby allow generalist pollinators to mediate successful reproduction in stressed plants. However, if all conspecifics lack resources of high enough quality for generalists, then these pollinators may disregard stressed plants altogether, leaving specialists as the sole floral visitors with the job of pollination. Reductions in the availability of high-quality floral resources may thus turn out to be detrimental for pollinators as well as their floral hosts. Plants species faced with drought may therefore increasingly trade-off allocation to separate, and perhaps competing, floral functions in order to attract a diversity of pollinators. Chapter 2, in part, is currently being prepared for submission for publication of the material. Gambel, Jess; Holway, David A. The dissertation author was the primary investigator and author of this material. The importance of day length for plant photosynthetic carbon gain has been widely recognized. Duration of the nocturnal period may also be important. During the nighttime, periods of minimal transpiration allow water potential gradients among plant parts to dissipate by internally redistributing water to tissues of lower water potential, including the movement of water to roots in dry soil layers . Nocturnal internal hydraulic redistribution may be very important to the maintenance of roots in surface soil layers where root densities are highest, as surface soil can be dry for extended periods in many regions . However, the rate at which water redistributes at night depends on a combination of environmental and plant factors, including the magnitude of differences in water potential among tissues and the hydraulic conductivity of the vascular system . Many factors influence the rates of tissue rehydration of roots at night. Nocturnal transpiration due to incomplete closure of stomata or high vapour pressure deficits can create and maintain high water potential gradients between the soil and leaves and thus limit water movement to other parts of the root system. In addition,rehydration of roots in dry soil layers may be slower where there are significant constraints to water movement, including embolisms, small-diameter and curved xylem vessels with extensive branching , high frequency of pits and end plate membranes , and the presence of heartwood, latewood, and rays . Lateral water movement may also be limited by the high resistances that form at the stem base–root junction . Vaccinium corymbosum has been reported to be a species that does not effectively distribute water laterally . Using a split-root water application, Abbott and Gough found that dyes did not move laterally from one stem to another and observed root mortality in the unwatered root container. The main objective of this study was to quantify patterns of nocturnal internal hydraulic redistribution and conductances among the first seven orders of V. corymbosum roots under severe drought conditions and its implications for root tissue rehydration.