Seed disperser effectiveness: The quantity component and patterns of seed rain for Prunus mahaleb

In this paper we analyze the seed dispersal stage of the Prunus mahaleb-fru givorous bird interaction from fruit removal through seed delivery within t he context of disperser effectiveness. The effectiveness of a frugivorous s pecies as a seed disperser is the contribution it makes to plant fitness. E ffectiveness depends on the quantity of seed dispersed ("quantity component ") and the quality of dispersal provided each seed ("quality component"). F or the main frugivores, we studied abundance, visitation rate, and feeding behavior, the major variables influencing the quantity component of effecti veness, and the postforaging microhabitat use and resultant seed shadows, w hich set the stage for postdispersal factors that will influence the qualit y component of effectiveness. Legitimate seed dispersers (SD) swallowed fruits whole and defecated or reg urgitated intact seeds; pulp consumers (PC) pecked fruits to obtain pulp an d dropped seeds to the ground, but some species occasionally dispersed seed s (PCSD species). Overall numbers of fruits removed (i.e., handled) by avia n frugivores were similar in the two study years; however, the estimated pe rcentage of seeds dispersed differed significantly, with lower relative dis persal success in the year with greater relative abundance of PC species. S imilar numbers of seeds were dispersed in the two years despite nearly a fo urfold difference in number of fruits produced. Fruit crop size explained > 80% variance in the number of seeds dispersed per tree. A total of 38 species of birds were recorded during censuses, with frugivor es representing 68.8% of them; the relative representation of SD, PC, and P CSD species was 42.2%, 17.2%, and 9.4%, respectively. Individual trees show ed extensive variation in visitation rates, ranging from 0.3 to 41.6 visits /10 h in any year. The main visitors were the SD species Phoenicurus ochrur os (10.8 visits/10 h), Turdus viscivorus (9.2 visits/10 h), Erithacus rubec ula (3.5 visits/10 h), and Sylvia communis (2.6 visits/10 h); and the PC sp ecies Fringilla coelebs (16.7 visits/10 h) and Pants ater (4.7 visits/10 h) . Species with large quantity components of effectiveness typically had eithe r high visit or high feeding rates, combined with high probability of dispe rsing a handled seed. Variation among species in fruit-handling behavior, h owever, was the main factor influencing variation in the quantity component . Visit rate in turn was influenced largely by local abundance. No single f rugivore trait, however, can adequately estimate the quantity component of disperser effectiveness. A "gulper''/''masher" dichotomy helps explain diff erences in fruit handling among major frugivore types and shows many correl ates with other aspects of frugivore activity that ultimately influence the quantity component. Most species showed marked preferences for microhabitats with plant cover, especially P. mahaleb, midheight shrubs, and Pinus (86.1% of the departure flights) and avoided open microhabitats. Most flights were over short dista nces (77.5% to perches located within 30 m). Among the main frugivores, 40. 3% of the exit flights were to perches >15 m away from the feeding tree, bu t only 18.5% of these flights were to perches >15 m from any P, mahaleb. Co vered microhabitats received significantly more seeds (39.3 +/- 5.0 seeds d ispersed/m(2), 1988 [mean +/- 1 SE]; 31.7 +/- 5.9 seeds dispersed/m(2), 198 9) than open microhabitats (2.8 +/- 0.7 seeds dispersed/m(2), 1988; 1.8 +/- 0.4 seeds dispersed/m2, 1989). The potential contribution of each bird species to the seed rain in each mi crohabitat was estimated from the number of visits recorded, the mean numbe r of seeds dispersed per visit, and the proportion of exit flights to each microhabitat. Microhabitats differed strongly in the proportions of seeds d elivered by the main frugivores, and bird species also differed in the prop ortions of seeds delivered to a given microhabitat. The seed rain to covere d microhabitats was delivered by a more heterogeneous assortment of species than the seed rain to open sites. The resulting seed shadow was a complex result,of the interaction between movement patterns of a suite of bird spec ies differing both in the quantity of seed dispersed and microhabitat prefe rences and in the landscape distribution of these microhabitat patches. Thi s seed shadow was extremely nonrandom due both to a strong overall preferen ce by most of the birds for the relatively scarce covered microhabitats and to species-specific preferences for particular types of covered microhabit ats. Different microhabitat types not only received variable amounts of dis persed seed, but also differed in the number and identity of disperser spec ies contributing to that seed rain.