Xiaojie Li’s research while affiliated with University of Kentucky and other places

The avifrugivore availability hypothesis predicts that summer-fruiting species will have an extended fruiting season and slow fruit removal, and the foliar flag hypothesis predicts that fruit dispersal in autumn-fruiting species coincides with fall bird migration. Phenology of fruit removal from plants of the early summer-fruiting shrub Rhus aromatica Ait. (var. aromatica) and of the later summer-fruiting shrub Rhus glabra L. (Anacardiaceae) was studied primarily to test the avifrugivore availability and foliar flag hypotheses of fruit dispersal, respectively. Fruits of R. aromatica ripened synchronously in early June, and 77-89% of them were gone from the plants by the early July, thus failing to support the avifrugivore availability hypothesis. The pattern of rapid fruit removal in this species, which was consistent throughout a 3-year period, is in contrast with previous reports of dispersal until September and even the next summer. Display size (at both infructescence and clump levels) had no effect on removal rate in R. aromatica. Fruits of R. glabra matured in August, and about 20% of them remained undispersed the following spring. In both species, fruit removal was mostly due to dispersers rather than to natural fruit fall. No correlations were found between fruit characteristics (pulp weight, pulp-to-seed ratio, moisture content of pulp) and probability of removal during the dispersal season in either species. Browsing by mammals, most likely white-tailed deer, was responsible for the rapid removal of fruits from R. aromatica plants during the 1st week after they turned mature-red; thereafter, most fruit removal was by birds. Two removal peaks were found for R. glabra fruits: a small one in September and a large one in winter to early spring. These peaks correspond to the fall migration season for migratory birds and the scarcity of food for winter-resident birds, in the study area, respectively. Thus, the dispersal pattern of R. glabra fruits does not support the foliar flag hypothesis.Key words: avifrugivore availability hypothesis, foliar flag hypothesis, fruit characteristics, fruit dispersal by birds or mammals, fruiting phenology, Rhus aromatica, Rhus glabra.

Soil seed bank dynamics, greenhouse seed germination and field seedling survival were compared in the weakly-clonal shrub Rhus aromatica Ait. and the strongly clonal shrub R. glabra L. (Anacardiaceae). Although both species had persistent soil seed banks, their sizes and dynamics differed greatly. The high seed density (5668 ± 1267 m−2 to 8739 ± 3303 m−2) in R. glabra was depleted by ca. 5% annually and replenished with 1102 ± 316 to 1828 ± 486 seeds m−2, whereas the low seed density (9 ± 5 to 138 ± 68 m−2) in R. aromatica was depleted by ca. 50% annually and replenished with 37 ± 7 to 69 ± 18 seeds m−2. Seeds of both species germinated within established populations. However, recruitment from seedlings was successful in R. aromatica but not in R. glabra, which depended exclusively on root suckering for aboveground population increase. This difference in local recruitment modes between the two species was correlated with differences in soil seed bank dynamics and other life history traits, such as fecundity, seed weight, seed dormancy, seed dispersal, seedling establishment and growth form. The results of this study have important implications about life history evolution in that they suggest that various traits related to increasing sexual reproduction in time and space evolved in concert with clonal reproduction in clonal woody species.

Abstract Physical dormancy (PY) is caused by a water-impermeable seed or fruit coat. It is known, or highly suspected, to occur in nine orders and 15 families of angiosperms (sensu Angiosperm Phylogeny Group 1998), 13 of which are core eudicots. The Zingiberales is the only monocot order, and Cannaceae (Canna) the only monocot family, in which PY is known to occur. Six of the nine orders, and 12 of the 15 families, in which PY occurs are rosids. Furthermore, six of the families belong to the Malvales. The water-impermeable palisade layer(s) of cells are located in the seed coats of 13 of the families, and in the fruit coats of Anacardiaceae and Nelumbonaceae. In all 15 families, a specialized structure is associated with the water-impermeable layer(s). The breaking of PY involves disruption or dislodgment of these structures, which act as environmental ‘signal detectors’ for germination. Representatives of the nine angiosperm orders in which PY occurs had evolved by the late Cretaceous or early Tertiary (Paleogene). Anatomical evidence for PY in fruits of the extinct species Rhus rooseae (Anacardiaceae, middle Eocene) suggests that PY had evolved by 43Ma, and probably much earlier. We have constructed a conceptual model for the evolution of PY, and of PY+ physiological dormancy (PD), within Anacardiaceae. The model begins in pre-Eocene times with an ancestral species that has large, pachychalazal, non-dormant (ND), recalcitrant seeds. By the middle Eocene, a derived species with relatively small, partial pachychalazal, orthodox seeds and a water-impermeable endocarp (thus PY) had evolved, and by the Oligocene, PD had been added to the seed (true seed + endocarp) dormancy mechanism. It is suggested that climatic drying (Eocene), followed by climatic cooling (Eocene–Oligocene transition), were the primary selective agents in the development of PY. An evolutionary connection between PY and recalcitrance is suggested by the relatively high concentration of these two character states in the rosids. Phylogenetic data and fossil evidence seem to support the PY→(PY+PD) evolutionary sequence in Anacardiaceae, which also may have occured in Leguminosae.

Anatomy of the endocarp was studied in relation to the physical dormancy-breaking mechanisms in experimentally treated Rhus aromatica var. aromatica and R. glabra germination units, which include seed plus endocarp (hereafter seeds). The endocarp has three distinct layers, with brachysclereids on the outside, osteosclereids in the middle, and macrosclereids on the inside. Brachysclereids in the carpellary micropyle region (i.e., region immediately adjacent to the integumentary micropyle) are shorter than those in other parts of the endocarp, and the macrosclereids in this region are not elongated. Thus, a weak point is formed in the endocarp. Concentrated sulfuric acid broke seed dormancy in R. aromatica by eroding the brachysclereids and osteosclereids in the carpellary micropyle region, whereas boiling water broke dormancy in seeds of R. glabra by inducing a blister adjacent to the carpellary micropyle.

Ontogeny and anatomy of the pericarp were studied in Rhus aromatica Ait. var. aromatica of subgenus Lobadium and in R. glabra L. of subgenus Rhus. The exocarp sensu stricto in both species originated from the outer epidermis of the ovary wall and remained single-layered. However, in a mature, desiccated fruit, the exocarp was united with the outer mesocarp, and this united-layer was physically detached from the rest of the fruit, thus forming a thin, papery peel. The vascular bundles together with the inner mesocarp, which remained parenchymatous in R. aromatica and became sclerified in R. glabra, was attached to the future stone in R. aromatica but detached from it in R. glabra. The endocarp sensu stricto of both species originated from the inner epidermis of the ovary wall. At anthesis, the proto-endocarp was 3-layered; one and two weeks after anthesis in R. glabra and R. aromatica, respectively, it had become distinctively 4-layered, and the innermost layer had elongated 5-10 times. Five weeks after anthesis in R. aromatica but only three weeks after anthesis in R. glabra, all layers except the outermost crystalliferous one were fully elongated and had begun to lignify. In both species, the crystalliferous layer remained parenchymatous, did not elongate, and contained crystals. The similarities in ontogeny and anatomy of the pericarp in these two species support the opinion that Lobadium should be kept as a subgenus within Rhus, rather than being elevated to the generic level. However, more taxa of Rhus need to be studied before the taxonomy of the genus can be clarified.

Morphology and physiology of fruit and seed development were compared in Rhus aromatica and R. glabra (Anacardiaceae), both of which produce drupes with water-impermeable endocarps. Phenology of flowering/fruiting of the two species at the study site was separated by ∼2 mo. However, they were similar in the timetable and pattern of fruit and seed development; it took ∼2 mo and ∼1.5 mo for flowers of Rhus aromatica and R. glabra, respectively, to develop into mature drupes. The single sigmoidal growth curve for increase in fruit size and in dry mass of these two species differs from the double-sigmoidal one described for typical commercial drupes such as peach and plum. Order of attainment of maximum size was fruit and endocarp (same time), seed coat, and embryo. By the time fruits turned red, the embryo had reached full size and become germinable; moisture content of seed plus endocarp had decreased to ∼40%. The endocarp was the last fruit component to reach physiological maturity, which coincided with development of its impermeability and a seed plus endocarp moisture content of <10%. At this time, ∼50, 37, and 13% of the dry mass of the drupe was allocated to the exocarp plus mesocarp unit, endocarp, and seed, respectively. The time course of fruit and seed development in these two species is much faster than that reported for other Anacardiaceae, including Rhus lancea, Protorhus, and Pistacia.

Fourteen seedlots of five species of Rhus were surveyed for presence/absence of physiological dormancy and/or for germination requirements of non-dormant seeds. Physiological dormancy was present in the four seedlots of R. aromatica studied, but not in either of the two seedlots of its close relative R. trilobata, which is in contrast to previous reports. Neither were seeds of R. glabra, R. typhina, nor R. virens physiologically dormant. Stratification at 5oC for 1 week or incubation in 500 or 1000 mg/l solutions of gibberellic acid broke physiological dormancy in > 90% of the R. aromaticaseeds. Maturation desiccation acted as a switch from a developmental to a germinative mode in R. aromatica embryos, whereas it was not required for germination of R. glabra or R. virens (R. trilobata and R. typhinanot tested). Seeds of all five species incubated on a moist substrate became fully imbibed in 2 d, at which time moisture content was approx. 70–80%of their initial weight. In general, germination of non-dormant seeds was rather insensitive to temperature and light. Seeds germinated equally well in light and in darkness over a daily (12 h/12 h) temperature range of 15/6–35/20oC. Over a 4 week period, the best germination percentages were obtained at 25/15 and 20/10oC, whereas 35/20oC appeared to be supraoptimal, though not always significantly so. If the incubation period was extended to 30 weeks, germination percentages were as high at 15/6oC as at 25/15 and 20/10oC.

Seed (= seed plus endocarp) morphology and physical dormancy were studied in seven North American Rhus species: R. copallina, R. glabra, R. typhina, R. aromatica (var. aromatica), R. microphylla, R. trilobata, and R. virens (var. virens). Seeds of Rhus glabra and R. typhina, of subgenus Rhus, were gray, approx. 3 mm long, 2 mm wide, and weighed approx. 7 mg, whereas those of R. aromatica, R. trilobata, and R. virens, of subgenus Lobadium, were brown, >4 mm long, approx. 4 mm wide, and weighed approx. 14.5 mg (R. trilobata) to approx. 23 mg (R. aromaticaand R. virens). Dormancy in all seeds was due to a water-impermeable endocarp, but depth of dormancy varied greatly among species and seedlots. After 4 weeks, 29–34% of the seeds from all five seedlots of R. trilobata, R. microphylla, and R. virens incubated on moist substrate had imbibed, compared to 0–14% of the seeds of all 16 seedlots of R. aromatica, R. copallina, R. glabra, and R. typhina. After 1 yr, imbibition among seedlots of R. aromaticavaried from 28 ± 2% to 69 ± 5%, whereas 93 ±2% to 100 ± 0% of the seeds from all seedlots of R. microphylla, R. trilobata, and R. virens did so. Neither dry laboratory storage for up to 4 yr (even 29 yr in R. aromatica) nor dry heating at 100oC or at 120oC effectively broke dormancy in any of the species tested (R. aromatica, R. glabra, R. trilobata, R. virens). Immersion in boiling water was the best method to render seeds of R. glabra and R. typhinapermeable, yet it was ineffective for those of R. aromatica, R. trilobata, and R. virens. In contrast, a 1 h-soaking in concentrated H2SO4 led to complete loss of endocarp impermeability in the latter three species, but mostly was ineffective in R. glabra and R. typhina. Thus, there seems to be a tendency for seeds of subgenus Rhusto respond well to boiling in water, but not to soaking in H2SO4, whereas the opposite is true for those of subgenus Lobadium.