| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | SEARCH RESULT |
|
|
|
|||||||
|
|||||||||
What's this? |
Reproductive Biology |
Department of Biology, Amherst College, Amherst, Massachusetts 01002 USA; Department of Ecology and Evolutionary Biology, University of Colorado, Boulder, Colorado 80309 USA
Received for publication March 30, 2007. Accepted for publication August 13, 2007.
ABSTRACT
Andromonoecy is hypothesized to evolve as a mechanism enabling plants to independently allocate resources to female and male function. If staminate flower production is a mechanism to regulate allocation to female function (i.e., fruit production), then large-fruited species should be more strongly andromonoecious than smaller-fruited taxa because more resources are required to mature large fruit. We combined phylogenetically independent contrast analyses with extensive phenotypic characterization under common greenhouse conditions to examine the predicted relationship between fruit mass and the strength of andromonoecy among 13 species in Solanum sections Acanthophora and Lasiocarpa. The strength of andromonoecy, defined as the proportion of staminate flowers produced within inflorescences, was significantly and positively associated with fruit mass in both naïve and phylogenetically independent analyses. Our results are consistent with the hypothesis that andromonoecy functions as a mechanism to regulate allocation to female function and suggest that the strength of andromonoecy is also associated with resource limitation. In general, we find that strong andromonoecy appears to arise via reductions in hermaphroditic flower number. However, increases in staminate flowers have also contributed to transitions to strong andromonoecy in certain species. Finally, our analyses identified a suite of correlated characters (flower size, ovary width, fruit mass) that are associated with changes in the sexual expression of andromonoecy.
Key Words: Acanthophora • andromonoecy • CAIC • flower size • fruit mass • fruit size • independent contrasts • Lasiocarpa • PDAP • phylogeny • Solanaceae • Solanum
Andromonoecy is a sexual system in which plants produce both hermaphroditic and female-sterile (staminate) flowers. Although the number of andromonoecious angiosperm species is relatively modest (approximately 4000 [Yampolsky and Yampolsky, 1922
]), these species are nested in at least 33 families (Yampolsky and Yampolsky, 1922
; Miller and Diggle, 2003
). This distribution suggests numerous independent origins of andromonoecy, and considerable attention has focused on identifying the conditions under which this sexual system has evolved and diversified (e.g., Bertin, 1982
; Whalen and Costich, 1986
; Anderson and Symon, 1989
; Spalik, 1991
; Diggle, 1993
, 1994
; Podolsky, 1993
; Emms, 1996
; Emms et al., 1997
; Elle and Meagher, 2000
; Miller and Diggle, 2003
; Vallejo-Marín and Rausher, 2007
).
A common theme in discussions of the evolution of andromonoecy is the suggestion that production of both hermaphroditic and staminate flowers allows resource allocation to female and male reproductive function to be flexible (Lloyd, 1980
; Bertin, 1982
; Solomon, 1985
; Sutherland, 1986
; Diggle, 1994
; Miller and Diggle, 2003
). Consistent with this idea, andromonoecy is often associated with resource limitation of fruit set and with individual fruits that are large and costly (see summary in Bertin, 1982
; Lloyd, 1979
; Primack and Lloyd, 1980
; Lloyd and Bawa, 1984
; Sutherland, 1986
; Whalen and Costich, 1986
; May and Spears, 1988
; Spalik, 1991
; Diggle, 1993
, 1994
; Emms, 1996
). Termination of gynoecial development before anthesis and the resulting production of a morphologically staminate flower prevent allocation to a fruit that cannot mature while maintaining potential male function (Ruiz Zapata and Kalin Arroyo, 1978
; Solomon, 1986
; Sutherland, 1986
; Spalik, 1991
).
Although these resource allocation hypotheses address the selective advantages that may drive the evolutionary origin of andromonoecy, they have not been used to explain subsequent evolutionary diversification in the degree of andromonoecy observed among species. Because of the modular nature of flower production, the expression of andromonoecy varies quantitatively; that is, individuals and species exist along a continuum from weakly andromonoecious (i.e., producing a small proportion of staminate flowers per inflorescence) to strongly andromonoecious (i.e., producing a large proportion of staminate flowers per inflorescence).
Andromonoecy is particularly common and variable in strength among members of the genus Solanum subgenus Leptostemonum.Whalen and Costich (1986)
proposed that variation in sexual expression among these species should be directly related to fruit size. They reasoned that if, as is commonly hypothesized, the production of staminate flowers via suppression of gynoecial development is a mechanism to control fruit initiation, then larger-fruited species should be more strongly andromonoecious than smaller-fruited species. Their analysis of fruit diameter and the strength of andromonoecy showed a strong correlation for members of Solanum sections Lasiocarpa and Acanthophora.
This correlation, however, was based on limited sampling within species and did not control for nonindependence of species due to common ancestry. Because traits like fruit size and sexual expression may be correlated with phylogeny, a more powerful test of the association should consider shared evolutionary history among these taxa. Further, it is not clear from this analysis how andromonoecy is actually related to fruit size. Because the strength of andromonoecy is expressed as a proportion of flower types, the observed correlation could be due to evolutionary changes in hermaphroditic or staminate flower number, or both. Moreover, changes in the proportion of flower types must occur within the context of total flower production, yet it is unknown whether changes in the strength of andromonoecy are necessarily associated with increases or decreases in flower number (Whalen and Costich, 1986
; Anderson and Symon, 1989
). Fruit size is also likely to be developmentally related to the size of the ovary from which the fruit develops (Primack, 1987
; Gillaspy et al., 1993
; Frary et al., 2000
), and ovary size, in turn, may be correlated to the sizes of other floral organs (Elle, 1998
; Ashman and Majetic, 2006
). Thus, evolutionary changes in fruit size may be correlated with changes in both ovary and flower size, and these may also show a relationship with evolutionary change in the strength of andromonoecy.
The relationship among the strength of andromonoecy and associated floral and fruit characters can be examined with comparative data coupled with phylogenetic hypotheses for relationships among species. The closely related members of Solanum sections Lasiocarpa and Acanthophora vary extensively in the expression of andromonoecy. Staminate flower production varies among species from near zero to 90% of flowers per inflorescence (Diggle, 1993
; Miller and Diggle, 2003
) (Table 1). In addition, hypotheses of evolutionary relationships for these groups are available (Bruneau et al., 1995
; Bohs, 2004
; Levin et al., 2006
), and the monophyly of sections Lasiocarpa (Bohs, 2004
) and Acanthophora sensu stricto (Levin et al., 2005
), as well as the sister relationship between the two clades, is well supported (Bohs, 2004
; Levin et al., 2005
).
|
MATERIALS AND METHODS
Study species
Section Lasiocarpa is a small (12 species) monophyletic section within the spiny Solanum group (subgenus Leptostemonum; Bohs, 2004
; Levin et al., 2006
). Most of the species occur in northwestern South America, although two species are found in Asia and the Pacific Islands. As originally circumscribed (Nee, 1979
), the approximately 20 species in section Acanthophora (also in subgenus Leptostemonum) are not monophyletic; however, the majority of species traditionally classified in this section and included to date in phylogenetic analyses (Levin et al., 2005
, 2006
) are in a strongly supported monophyletic group (see Levin et al., 2006
) that includes the five species of Acanthophora reported on here. Further, there is strong support for a sister relationship between sections Lasiocarpa and Acanthophora (Levin et al., 2005
, 2006
). Species of both sections are all sexually reproducing, self-compatible, and andromonoecious (Nee, 1979
; Whalen et al., 1981
).
All species were cultivated in greenhouses at the University of Colorado (Boulder, Colorado, USA). The study included 13 species: eight of 12 species in section Lasiocarpa, including Solanum candidum, S. lasiocarpum, S. repandum, S. hirtum, S. pectinatum, S. pseudolulo, S. quitoense, and S. stramoniifolium; and five of the 19 species in the Acanthophora clade (as defined in Levin et al., 2006
), including S. acerifolium, S. capsicoides, S. mammosum, S. palinacanthum, and S. tenuispinum.
Cultivation of plants and data collection
Plants were grown from seed, and 5–12 genotypes for each species were clonally replicated via vegetative cuttings to produce genetically identical replicates. Each clonal replicate was transplanted into an 11-L pot containing a 2 : 1 mix of Fafard Growing Mix #2 (Conrad Fafard, Agawam, Massachusetts, USA) to Persolite (Persolite Products, Florence, Colorado, USA) plus Osmocote 13-13-13 slow-release fertilizer (Scotts, Marysville, Ohio, USA). Plants were watered daily with 150–200 ppm of Excel Magnitrate fertilizer (Scotts).
Clonal replicates for each genotype were randomly assigned positions in the greenhouses. We pollinated all open hermaphroditic flowers every other day using a mixture of pollen collected from several (three or more genotypes) conspecific pollen donors. Hermaphroditic flowers remained open for 2 to 3 d, and most flowers were pollinated at least twice.
Two branches on each plant were selected and censused every other day. Early developing, basal inflorescences were numbered first, and floral positions within inflorescences were also numbered from the base to the tip. Inflorescence position, flower position, floral sexual phenotype (i.e., hermaphrodite or staminate), and fruit production were recorded for all flowers on up to 10 inflorescences on each of two marked branches per individual. Data were summarized initially for each inflorescence position on each genotype for each species. We then calculated the average number of hermaphroditic, staminate, and total flowers per inflorescence, as well as the average number of fruits per inflorescence for each species. The strength of andromonoecy was expressed as the proportion of staminate flowers produced within inflorescences. Proportions were arcsine-square-root transformed (Sokal and Rohlf, 1995
) prior to analyses and were back transformed for presentation.
All fruits produced on marked branches were collected, dried completely at 60°C, and weighed to the nearest milligram on a XL-400D balance (Denver Instruments, Denver, Colorado, USA). In total, 7310 fruit were weighed across 12 (excluding S. acerifolium) species (range 28–2197 fruits per species), and average fruit mass for inflorescence, genotype, and species combinations was calculated. Fruit mass was multiplied by 100 and log transformed prior to analyses.
Flowers from uncensused branches were collected for floral measurements. Because floral position within an inflorescence strongly affects the size of structures (Diggle and Miller, 2004
), we used only hermaphroditic flowers produced in basal positions within inflorescences. Recently opened flowers (<48 h old) were collected and fixed in FAA (formalin-acetic acid-alcohol; Berlyn and Miksche, 1976
) for 24 h before transfer to 70% ethanol for storage. Flowers were measured using digital calipers (for large corollas) and a Zeiss (Thornwood, New York, USA) Stemi SV-11 dissecting microscope equipped with a Zeiss Axiophot digital camera and image analysis system. For each flower, eight measurements were made: length and width of the dorsal petal, anther length and width, style length, stigma width, and ovary length and width. Principal component analysis was used to summarize variation in the floral characters (JMP version 5.0.1, SAS Institute, 1989–2002
). The first principal component was used as an index of flower size in the contrast analyses. In total, we measured 585 flowers collected from 101 genotypes from 12 species (all except S. acerifolium). We measured 5.8 flowers, on average, from within genotypes and included an average of 8.4 genotypes for each species. All floral measurements were log transformed prior to analyses.
Interspecific correlations
Interspecific correlations among characters were calculated as Pearson product–moment correlations using the species means for each of the characters. Data summaries and correlation analyses were done in JMP version 5.0.1 (SAS Institute, 1989–2002
).
Phylogenetic hypotheses and independent contrasts
Several hypotheses of evolutionary relationships were used to conduct independent contrasts analyses. First, we used results from Bohs (2004)
, who analyzed chloroplast-sequence data from the trnT-trnL and trnL-trnF spacers and the trnL gene to infer relationships within section Lasiocarpa. That study also included the five members of section Acanthophora studied here. We used relationships depicted in Fig. 1 of Bohs (2004)
with one modification. In that topology, two accessions of S. repandum were not monophyletic, and here we used two trees, each depicting one of the positions for this taxon (topologies IA and IB in Fig. 1).
|
, who inferred relationships among a larger set of taxa in Solanum subgenus Leptostemonum using a combined analysis of the nuclear internal transcribed spacer region, the granule-bound starch synthase gene, and the chloroplast spacer region trnS-trnG. Specifically, we used the topology from their Bayesian analysis of combined data (Fig. 3 in Levin et al., 2006
). There is one polytomy in this tree and weak support for the node uniting S. hirtum and S. pseudolulo (see topology II, Fig. 1); we reconciled the polytomy and relationships among S. hirtum, S. pseudolulo, and S. pectinatum in all possible ways and used these topologies (topologies IIA1–IIA3, IIB1–IIB3, and IIC1–IIC3 in Fig. 1) in independent contrasts.
Finally, we obtained sequence data for the trnS-trnG, trnT-trnF, ITS, and waxy regions from the authors of the published sources (Bohs, 2004
; Levin et al., 2005
, 2006
) and aligned these regions manually using SeAl (Rambaut, 2002
). Phylogenetic analysis of sequence data was carried out using maximum likelihood as implemented in PAUP* version 4.0b10 (Swofford, 2002
). Maximum likelihood model parameters were determined using the Akaike information criterion in Modeltest v. 3.7 (Posada and Crandall, 1998
). The best model (GTR+I+G) was used in a likelihood analysis in PAUP* using the heuristic search option, eight starting trees from a parsimony analysis, tree-bisection-reconnection (TBR) branch swapping, and the MulTrees option in effect. The five species in section Acanthophora were designated as outgroups in this analysis. We used the most likely tree from this analysis in independent contrasts (topology III in Fig. 1).
Phylogenetically independent contrasts were conducted for each of the 13 topologies depicted in Fig. 1 using the PDAP module version 1.07 (Midford et al., 2003
) in Mesquite version 1.06 (Maddison and Maddison, 2005
). All branch lengths were set to one (Diaz-Uriarte and Garland, 1998
). Diagnostic tests in PDAP were run to ensure that the contrasts were properly standardized (Garland et al., 1992
; Purvis and Rambaut, 1995
; Freckleton, 2000
). We tested the relationship between variables by least-squares linear regressions of the contrasts forced through the origin (Harvey and Pagel, 1991
; Garland et al., 1992
). Significant contrast correlations would support the hypothesis that evolutionary change in one character is consistently related (positively or negatively) to change in the second character across the clade.
Correlations between the strength of andromonoecy, defined as the proportion of staminate flowers produced within inflorescences, and the variables that make up that proportion (staminate, hermaphroditic, and total flower number) are inherently interrelated. We therefore explored the relationship of the strength of andromonoecy and changes in flower number using two alternative approaches. First, we used the BRUNCH procedure of the program CAIC (Purvis and Rambaut, 1995
) to examine correlated changes in a categorical variable and a continuous variable. Species were categorized as having either weak andromonoecy (20% or fewer staminate flowers within inflorescences) or strong andromonoecy (40% or higher staminate flowers within inflorescences), and we tested whether the evolution of continuous variables (total, hermaphroditic, or staminate flower production) is associated with the evolution of the categorical variable (weak vs. strong andromonoecy). Positive contrasts for the strength of andromonoecy would indicate that the continuous variable changed in the same direction as the categorical variable, whereas negative contrasts would indicate that the continuous variable changed in the opposite direction as the categorical variable.
In addition, we reconstructed the evolutionary shifts in flower number across the transition from weak to strong andromonoecy for selected taxa. Solanum quitoense and S. pectinatum are both strongly andromonoecious, yet they differ dramatically in both total flower production and the numbers of hermaphroditic and staminate flowers (Table 1). Whereas S. quitoense produces numerous flowers within inflorescences and many of these are staminate, S. pectinatum typically bears 2–3 flowered inflorescences with only a single hermaphroditic flower in each. Values for total, hermaphroditic, and staminate flower number were inferred for ancestral species using squared-change parsimony as implemented in Mesquite version 1.06 (Maddison and Maddison, 2005
). We calculated the change in flower number from the ancestor leading to S. quitoense and S. pectinatum to explore the evolutionary transitions of flower number associated with strong andromonoecy in each of these taxa. Reconstructions were repeated for all topologies as a test of the sensitivity of our inferences to uncertainty in phylogeny.
RESULTS
Species means for the strength of andromonoecy (proportion staminate flowers); fruit mass and number; total-, hermaphroditic-, and staminate-flower production; ovary width; and the first principal component are in Table 1. There was considerable variation among species for these characters. For example, fruit dry mass varied among species by more than an order of magnitude, from 0.33 to 7.34 g, and similarly the proportion of staminate flowers per inflorescence ranged from nearly zero to just over 0.90. The first principal component (PC1) accounted for 59% of the total variation in the data set and described overall flower size. The eigenvector associated with PC1 had positive loadings for all the floral characters, and these loadings were of similar magnitude, characteristic of a size vector.
Fruit mass, fruit number, and the strength of andromonoecy
As fruits got larger, the strength of andromonoecy increased among the Solanum species included in the interspecific analysis (Pearson product moment correlation, r = 0.65, P = 0.022; Table 2; Fig. 2). Further, this relationship persisted after accounting for evolutionary history using independent contrasts. Contrast correlations involving fruit mass and the strength of andromonoecy were positive and significant (0.58 
r 0.66, P 0.05; Table 2) for all but three topologies in which S. quitoense is basal within Lasiocarpa (topologies IIA1–IIA3, see Fig. 1). The relationship was marginally significant even for these topologies (0.54 r 0.57, 0.05 < P 0.07; Table 2).
|
|
r –0.62, 0.008 P 0.031; Table 2).
Flower number and the strength of andromonoecy
Interspecies correlations of the strength of andromonoecy and total, hermaphroditic, and staminate flower numbers are inherently interrelated because the latter are used to calculate the strength of andromonoecy. Of the variables used to calculate the strength of andromonoecy, hermaphroditic flower number was negatively associated with fruit mass (r = –0.66, P = 0.02), but no relationship was detected between fruit mass and either total (r = –0.43, P = 0.16) or staminate (r = 0.52, P = 0.08) flower number. This pattern remained following contrast correlations; hermaphroditic flower number was negatively related to fruit mass for all topologies (–0.67 r –0.58, P 0.05), whereas correlations of fruit mass with either total or staminate flower number were not significant for any topology.
We also investigated the relationship between the strength of andromonoecy and flower production by analyzing the strength of andromonoecy as a categorical variable. Because of small sample sizes (i.e., three to five transitions between weak and strong andromonoecy within topologies), testing for significance was limited. Despite this limitation, results of the categorical analysis were similar to those contrasts based on the continuous variables described in the previous paragraph. No relationship of andromonoecy to total or staminate flower production was detected, whereas contrasts for hermaphroditic flower number were negative for all topologies and significant for four of these (topologies IIA2–3 and IIB2–B3; 13.5 F 221.9, 0.005 P 0.035).
Finally, we examined evolutionary transitions in flower number using ancestral character state reconstructions. We focused on the species S. quitoense and S. pectinatum, because these taxa differ widely in total flower number per inflorescence (9.5 and 2.6, respectively; Table 1) while sharing strong andromonoecy (Table 1). Figure 3 shows the change in hermaphroditic and staminate flower number from the relevant ancestral species to S. quitoense and S. pectinatum. Hermaphroditic flower number decreased for both species in all topologies (Fig. 3) and significantly more so for some topologies in S. pectinatum. In contrast, staminate flower number remained relatively unchanged in S. pectinatum, whereas staminate flower number increased in S. quitoense.
|
r 0.86, P 0.0008; flower size and fruit mass, 0.77 r 0.85, P 0.003; flower size and ovary width, 0.68 r 0.84, P 0.015). In addition to being correlated with one another, ovary width and flower size were also significantly correlated with the strength of andromonoecy in both interspecific (Fig. 4D, E) and independent contrast analyses (ovary width and andromonoecy, 0.61 r 0.76, P 0.04; flower size and andromonoecy, 0.75 r 0.80, P 0.005).
|
Andromonoecy is thought to evolve as a mechanism that curtails allocation to female function within some flowers without altering male function (reviewed in Bertin, 1982
; Diggle, 1994
). Whalen and Costich (1986)
extended this independent allocation hypothesis in an attempt to explain the broad variation in the strength of andromonoecy observed among Solanum species. They reasoned that if staminate flower production is a mechanism to regulate allocation to female function (fruit production), then large-fruited species should be more strongly andromonoecious than smaller-fruited taxa because more resources are required to mature large fruit. Our investigation combined phylogenetically independent contrasts with extensive phenotypic characterization to examine the predicted relationship between fruit mass and the strength of andromonoecy in Solanum. We confirmed this prediction and showed that the correlation between fruit size and the strength of andromonoecy is due largely to an underlying association between evolutionary changes in fruit mass and hermaphroditic flower number. Moreover, fruit mass, ovary size, and flower size comprise a suite of characters that evolve with changes in the expression of andromonoecy.
Evolution of fruit size and the strength of andromonoecy
As predicted (Whalen and Costich, 1986
), the interspecific correlation of fruit mass and the strength of andromonoecy is positive and significant (Table 2; Fig. 2). More importantly, this relationship persists after accounting for shared evolutionary history and is generally robust to uncertainty in evolutionary relationships (Table 2; Fig. 1). Many of the contrast correlation coefficients are very close to the interspecific values, suggesting that there is little phylogenetic constraint on these characters (Armbruster et al., 2002
). Fruit characteristics are sometimes considered to be evolutionarily conservative (Spujt, 1994
; Knapp, 2003
), yet within this small group of Solanum, fruit size is apparently quite labile. The ecology of frugivory and fruit dispersal in relation to fruit size in this group merits investigation.
Both interspecific and independent contrast correlations confirm that, within the clade containing Solanum sections Acanthophora and Lasiocarpa, evolutionary changes in the strength of andromonoecy generally are positively associated with changes in fruit size. Particular contrasts, however, are not consistent with this pattern. The contrast between S. repandum and S. pseudolulo in topology IA (Fig. 1) and the contrast of and S. mammosum and S. palinacanthum in many topologies (topologies, II–III; Fig. 1) represent cases in which weaker andromonoecy is associated with larger fruit. The species pair, S. mammosum and S. palinacanthum, both have strong andromonoecy (>60% staminate flower production within inflorescences; Table 1) and the largest fruits of any Acanthophora (Levin et al., 2005
). Because the sister relationship of these species is strongly supported in most topologies (Fig. 1) and also in a more comprehensive phylogenetic investigation of section Acanthophora (Levin et al., 2005
), it is likely that the evolution of both large fruits and strong andromonoecy occurred in the common ancestor of these taxa. Subsequent evolution of fruit of these species, however, is likely related to other aspects of their reproductive biology. For example, S. mammosum produces very unusual, elongated fruits that are subtended by supernumerary carpels (termed mammillae; Miller, 1969
). Despite the presence of large fruit and relatively weak andromonoecy in S. mammosum (as compared with S. palinacanthum), within each species, the genotype means of the strength of andromonoecy and of fruit mass are positively correlated (S. mammosum, N = 12 genotypes, r = 0.70, P = 0.011; S. palinacanthum, N = 8 genotypes, r = 0.82, P = 0.013). Thus, for both species, genotypes with large fruit are more strongly andromonoecious, a result consistent with the general relationship observed among species of Lasiocarpa and Acanthophora.
The prediction that strong andromonoecy is associated with large fruit (Whalen and Costich, 1986
) assumes that resources limit fruit production. Our data indicate trade-offs between fruit size and number for both interspecific and independent contrast analyses (Table 2); among the species studied, evolutionary changes in fruit size generally have been accompanied by inverse changes in fruit number. Such trade-offs between fruit size and number are common among fleshy-fruited taxa and are often attributed to resource competition and reallocation within inflorescences (Veliath and Ferguson, 1972
; Stephenson, 1981
; Wyatt, 1982
; Diggle, 1995
; Elle, 1996
; Lee, 1988
; Nesbitt and Tanksley, 2001
; Baldet et al., 2006
). It is unknown whether the among-species variation in fruit number seen here (Table 1) is a direct consequence of changes in fruit size (and resultant changes in resource competition within inflorescences) or whether separate regulation of fruit number also has evolved.
Flower number and the strength of andromonoecy
The relationship between evolutionary changes in fruit size and changes in the strength of andromonoecy must be due to an underlying correlation of fruit mass with the individual elements used to calculate the strength of andromonoecy: staminate flower number, hermaphroditic flower number, or both (total flower number). Neither total nor staminate flower number is correlated with fruit mass; a significant relationship to fruit mass holds only for hermaphroditic flower production. Thus, the relationship between strong andromonoecy and large fruit is due to an underlying association of fruit size and hermaphroditic flower number. There is no consistent relationship between evolutionary changes in the strength of andromonoecy and changes in staminate or total flower production. This conclusion is supported by contrast correlations in which the strength of andromonoecy is expressed as a categorical variable. Evolutionary changes in the strength of andromonoecy are related to variation in hermaphroditic flower number but not to staminate or total flower number.
Ancestral reconstructions of flower numbers also are consistent with the interspecific and contrast correlation analyses. Evolutionary transitions to strong andromonoecy are associated with a decrease in the number of hermaphroditic flowers per inflorescence, whereas changes in staminate flower number are variable. For S. pectinatum (Fig. 3), for example, the evolution of strong andromonoecy involved only decreases in hermaphroditic flower number with staminate flower number remaining relatively unchanged. In contrast, for S. quitoense (Fig. 3), there was both a decrease in hermaphroditic and an increase in staminate flower number. Thus, while changes in staminate flower number are not consistently related to evolutionary changes in the strength of andromonoecy across the clade as a whole, there are particular instances where staminate flower production clearly plays a role in modulating the strength of andromonoecy. Evolutionary changes in staminate flower number can affect the strength of andromonoecy, and the evolution of this character should be studied from multiple perspectives, not just in relation to fruit size. Staminate flower production may be related to pollinator attraction (Janzen, 1977
; Podolsky, 1993
; Elle and Meagher, 2000
; Connolly and Anderson, 2003
), pollen donation (Primack and Lloyd, 1980
; Harder and Thomson, 1989
; Harder and Barrett, 1996
; Elle and Meagher, 2000
; Harder et al., 2000
; Connolly and Anderson, 2003
), and even herbivore avoidance (Bertin, 1982
). Because the evolution of S. quitoense appears to have involved the addition of staminate flowers, this species would be a logical choice for further investigation of the fitness consequences of staminate flower production.
Production of staminate flowers, with their reduced ovary, is expected to result in recovery of resources that can be allocated to other, potentially fitness-enhancing functions (Primack and Lloyd, 1980
; Bertin, 1982
; Solomon, 1986
; Spalik, 1991
; Emms, 1993
; reviewed in Vallejo-Marín and Rausher, 2007
). For species of Lasiocarpa and Acanthophora, the lack of association between staminate flower production and the strength of andromonoecy suggests that although strongly andromonoecious taxa make fewer hermaphroditic flowers (fewer flowers with fully developed ovaries), they do not consistently reallocate saved resources into increased staminate flower production. Similarly, Vallejo-Marín and Rausher (2007)
found that resources recovered by not producing hermaphroditic flowers were not reallocated to increased staminate flower production, increased seed production, or to vegetative growth in andromonoecious S. carolinense, also a member of Solanum subgenus Leptostemonum.
Total flower production per inflorescence varies substantially among species of Lasiocarpa and Acanthophora (Table 2; Nee, 1979
; Whalen et al., 1981
). In their discussion of the diversification of andromonoecy, Whalen and Costich (1986)
observed that, among New World taxa (primarily sections Lasiocarpa and Acanthophora), strongly andromonoecious species tended to produce fewer flowers per inflorescence than their more weakly andromonoecious relatives and suggested that increased strength of andromonoecy has been accompanied by a loss rather than a gain of flowers within inflorescences in this group. In contrast, the evolution of andromonoecy among Solanum in Australia may have involved the addition of staminate flowers and increases in total flower number (Anderson and Symon, 1989
). We find no support for a regular association of either increases or decreases in total flower number with evolutionary changes in the strength of andromonoecy or with changes in fruit size. Although flower number per inflorescence appears to be an evolutionary labile character (it ranges from a mean of 2.6 in S. pectinatum to 16.5 in S. stramoniifolium), these differences are related to factors other than fruit size. Total flower number has been shown to affect pollinator attraction generally (Campbell, 1989
; Ehrlen, 1991
; Emms, 1993
; Harder and Barrett, 1996
) and female fitness in S. carolinense (Elle, 1999
).
Correlation does not demonstrate causation, but the observed association between strong andromonoecy, large fruits, and fewer hermaphroditic flowers is consistent with the hypothesis that andromonoecy is a mechanism of pre-anthesis regulation of fruit set (Bertin, 1982
; Whalen and Costich, 1986
). As fruit become larger, fewer fruit can be matured, and fewer flowers with functional ovaries are produced. Fewer hermaphroditic flowers, in the absence of any consistent decreases in total flower number, results in stronger andromonoecy. Support for the fruit regulation hypothesis is all the more compelling because recent experimental studies have failed to find support for alternative hypotheses for the production of staminate flowers in S. carolinense. As noted, resources saved by the production of staminate (as opposed to hermaphroditic) flowers are not reallocated to fitness-enhancing functions. The same study also found that staminate flowers do not have increased pollen donation relative to hermaphroditic flowers (Vallejo-Marín and Rausher, 2007
).
Evolution of a suite of correlated characters
Evolutionary changes in fruit mass, ovary size, and flower size are all positively correlated with one another in both interspecific and contrast analyses. That is, species with large flowers have large ovaries that develop into large fruit (Fig. 4). Genetic studies of tomato, Solanum lycopersicum, and its wild relatives provide insight into the correlation of ovary and fruit size. A single quantitative trait locus (QTL), fw2.2, has a major effect on fruit mass in tomato; allelic differences in fw2.2 increase fruit mass as much as 30% (Grandillo and Tanksley, 1996
; Cong et al., 2002
; Frary et al., 2002; Cong and Tanksley, 2006
). Mature fruit size in tomato is determined by mitotic activity in both the pre-anthetic ovary and the developing fruit (Bohner and Bangerth, 1988
; Gillaspy et al., 1993
; Joubés et al., 1999
; Cong et al., 2002
; Bertin, 2004
; Baldet et al., 2006
), and fw2.2 is associated with modulation of mitotic activity at both of these stages (Frary et al., 2002). Thus, variation in fw2.2 or a similar QTL will result in changes to both ovary and fruit size. The tomato locus fw2.2 is likely orthologous to the loci affecting fruit size in eggplant (Solanum melongena) and pepper (Capsicum annuum; Tanksley, 2004
). Perhaps this locus, or a similar gene of large effect, may explain the differences in fruit size among Solanum species in sections Lasiocarpa and Acanthophora and underlie the correlated changes in fruit and ovary size.
Although the effect of fw2.2 does not appear to extend to flower size, the joint evolution of flower and ovary size is not unexpected given that genetic correlations and developmental coregulation of floral organs is common (Elle, 1998
; Armbruster et al., 1999
; Ashman, 1999
; Caruso, 2004
; Anderson and Busch, 2006
; reviewed in Ashman and Majetic, 2006
). For example, analysis of andromonoecious S. carolinense revealed strong genetic correlations among corolla diameter, pistil length (ovary plus style and stigma), and anther length and width (Elle, 1998
). Such genetic correlations are predicted to underlie the phenotypic correlations of floral organs observed among species of Solanum sections Lasiocarpa and Acanthophora.
Fruit, flower, and ovary size are not only correlated with one another, they are each correlated with the strength of andromonoecy. Although the observed relationship between fruit size and the strength of andromonoecy supports the conclusion that evolutionary changes in fruit size are generally accompanied by changes in the strength of andromonoecy, the cause of this relationship remains unknown. Because fruit size is related to ovary and flower size and also is likely related to other, unmeasured, variables (Harvey and Pagel, 1991
; Price, 1997
), any of these may be the targets of selection. For example, pollinators generally tend to prefer larger flowers (Bell, 1985
; Conner and Rush, 1996
; Galen and Newport, 1987
; Stanton and Preston,
1988), and selection for increased flower size could, indirectly, result in changes in both fruit mass and the strength of andromonoecy.
Summary
Among the 13 species of Solanum studied here, fruit size is positively correlated with the strength of andromonoecy. This result persists after controlling for evolutionary history and is robust to uncertainty in phylogenetic relationships. In general, evolutionary changes in the strength of andromonoecy among members of sections Acanthophora and Lasiocarpa are associated with reductions in hermaphroditic flower production but not consistently with changes in staminate flower number. However, increases in staminate flower production have clearly led to strong andromonoecy in at least one species, S. quitoense. The association between large fruits, strong andromonoecy, and few hermaphroditic flowers is consistent with the hypothesis that andromonoecy provides a mechanism of pre-anthetic regulation of fruit set. The relationship of fruit mass to evolutionary changes in the strength of andromonoecy, however, cannot be considered in isolation. Variation in fruit size is correlated with changes in both ovary and flower size, and each is correlated with the strength of andromonoecy. Any of these, or some other unmeasured variable, could be the target of selection underlying the diversification of andromonoecy. Indeed, discerning the causal relationships among such highly correlated traits remains a challenge to our understanding of reproductive evolution, especially in wild species.
FOOTNOTES
1 The authors thank C. Heiser, L. Bohs, and the Botanical Garden of Nijmegen for seed accessions; T. Lemieux for advice and support regarding cultivation of plants; M. Allsop, E. Comfort, K. Ebert, M. Van Etten, T. Forbis, R. Khorsand, C. McGraw, N. Reynolds, K. Ryerson, and C. States for assistance with data collection; the Department of Ecology and Evolutionary Biology at the University of Colorado for greenhouse space; and W. Friedman for comments on earlier drafts of the manuscript. This work was supported by National Science Foundation grant DEB 9982489 to P.K.D. and NSF grant DEB 0343732 and an Amherst College Trustee Faculty Fellowship to J.S.M. 
4 Author for correspondence (pamela.diggle{at}colorado.edu
)
101 5 Authors contributed equally to this research.
LITERATURE CITED
Anderson G. J. Symon D. E.. 1989. Functional dioecy and andromonoecy in Solanum. Evolution 43: 204-219..[CrossRef][Web of Science]
Anderson I. A. Busch J. W.. 2006. Relaxed pollinator-mediated selection weakens floral integration in self-compatible taxa of Leavenworthia (Brassicaceae). American Journal of Botany 93: 860-867..
Armbruster W. S. Di Stillio V. S. Tuxill J. D. Flores T. C. Velásquez Runk J. L.. 1999. Covariance and decoupling of floral and vegetative traits in nine neotropical plants: a re-evaluation of Berg's correlation-pleiades concept. American Journal of Botany 86: 39-55..
Armbruster W. S. Mulder C. P. H. Baldwin B. G. Kalisz S. Wessa B. Nute H.. 2002. Comparative analysis of late floral development and mating-system evolution in tribe Collinsieae (Scrophulariaceae s.l.). American Journal of Botany 89: 37-49..
Ashman T.-L.. 1999. Quantitative genetics of floral traits in a gynodioecious wild strawberry, Fragaria virginiana: implications for the independent evolution of female and hermaphrodite floral phenotypes. Heredity 83: 733-741..[CrossRef][Web of Science][Medline]
Ashman T.-L. Majetic C. J.. 2006. Genetic constraints on floral evolution: a review and evaluation of patterns. Heredity 96: 343-352..[CrossRef][Web of Science][Medline]
Baldet P. Hernould M. Laporte F. Mounet F. Just D. Mouras A. Chevalier C. Rothan C.. 2006. The expression of cell proliferation-related genes in early developing flowers is affected by a fruit load reduction in tomato fruits. Journal of Experimental Botany 57: 961-970..
Bell G.. 1985. On the function of flowers. Proceedings of the Royal Society of London, B, Biological Sciences 224: 223-265..
Berlyn G. P. Miksche J. P.. 1976. Botanical microtechnique and cytochemistry. Iowa State University Press, Ames, Iowa, USA..
Bertin N.. 2004. Analysis of the tomato fruit growth response to temperature and plant fruit load in relation to cell division, cell expansion and DNA endoreduplication. Annals of Botany 95: 439-447..[CrossRef][Medline]
Bertin R. I.. 1982. The evolution and maintenance of andromonoecy. Evolutionary Theory 6: 25-32..
Bohner J. Bangerth F.. 1988. Effects of fruit set sequence and defoliation on cell number, cell size and hormone levels of tomato fruit (Lycopersicon esculentum Mill.) within a truss. Plant Growth Regulation 7: 141-155..[Web of Science]
Bohs L.. 2004. A chloroplast DNA phylogeny of Solanum section Lasiocarpa. Systematic Botany 29: 177-187..[CrossRef][Web of Science]
Bruneau A. Dickson E. E. Knapp S.. 1995. Congruence of chloroplast DNA restriction site characters with morphological and isozyme data in Solanum sect. Lasiocarpa. Canadian Journal of Botany 73: 1151-1167..[CrossRef]
Campbell D. R.. 1989. Inflorescence size: a test of the male function hypothesis. American Journal of Botany 76: 730-738..[CrossRef][Web of Science]
Caruso C. M.. 2004. The quantitative genetics of floral trait variation in Lobelia: potential constraints on adaptive evolution. Evolution 58: 732-740..[Web of Science][Medline]
Cong B. Liu J. Tanksley S. D.. 2002. Natural alleles at a tomato fruit size quantitative trait locus differ by heterochronic regulatory mutations. Proceedings of the National Academy of Sciences, USA 99: 13606-13611..
Cong B. Tanksley S. D.. 2006. fw2.2 and cell cycle control in developing tomato fruit: a possible example of gene co-option in the evolution of a novel organ. Plant Molecular Biology 62: 867-880..[CrossRef][Web of Science][Medline]
Conner J. K. Rush S.. 1996. Effects of flower size and number on pollinator visitation to wild radish, Raphanus raphanistrum. Oecologia 105: 509-516..[CrossRef][Web of Science]
Connolly B. A. Anderson G. J.. 2003. Functional significance of the androecium in staminate and hermaphroditic flowers of Solanum carolinense (Solanaceae). Plant Systematics and Evolution 240: 235-243..[CrossRef][Web of Science]
Diaz-Uriarte R. Garland T. Jr. 1998. Effects of branch length errors on the performance of phylogenetically independent contrasts. Systematic Biology 47: 654-672..[CrossRef][Web of Science][Medline]
Diggle P. K.. 1993. Developmental plasticity, genetic variation, and the evolution of andromonoecy in Solanum hirtum (Solanaceae). American Journal of Botany 80: 967-973..[CrossRef][Web of Science]
Diggle P. K.. 1994. The expression of andromonoecy in Solanum hirtum (Solanaceae): phenotypic plasticity and ontogenetic contingency. American Journal of Botany 81: 1354-1365..[CrossRef][Web of Science]
Diggle P. K.. 1995. Architectural effects and the interpretation of patterns of fruit and seed development. Annual Review of Ecology and Systematics 26: 531-552..[CrossRef][Web of Science]
Diggle P. K. Miller J. S.. 2004. Architectural effects mimic floral sexual dimorphism in Solanum (Solanaceae). American Journal of Botany 91: 2130-2040..
Ehrlen J.. 1991. Why do plants produce surplus flowers? A reserve-ovary model. American Naturalist 138: 918-933..[CrossRef][Web of Science]
Elle E.. 1996. Reproductive trade-offs in genetically distinct clones of Vaccinium macrocarpon, the American cranberry. Oecologia 107: 61-70..[CrossRef][Web of Science]
Elle E.. 1998. The quantitative genetics of sex allocation in the andromonoecious perennial, Solanum carolinense (L.) [sic]. Heredity 80: 481-488..[CrossRef][Web of Science]
Elle E.. 1999. Sex allocation and reproductive success in the andromonoecious perennial Solanum carolinense (Solanaceae). I. Female success. American Journal of Botany 86: 278-286..
Elle E. Meagher T. R.. 2000. Sex allocation and reproductive success in the andromonoecious perennial Solanum carolinense (Solanaceae). II. Paternity and functional gender. American Naturalist 156: 622-636..[CrossRef][Web of Science]
Emms S. K.. 1993. Andromonoecy in Ziadenus paniculatus (Liliaceae): spatial and temporal patterns of sex allocation. American Journal of Botany 80: 914-923..[CrossRef][Web of Science]
Emms S. K.. 1996. Temporal patterns of seed set and decelerating fitness returns on female allocation in Zigadenus paniculatus (Liliaceae), an andromonoecious lily. American Journal of Botany 83: 304-315..[CrossRef][Web of Science]
Emms S. K. Stratton D. A. Snow A. A.. 1997. The effect of inflorescence size on male fitness: experimental tests in the andromonoecious lily, Zigadenus paniculatus. Evolution 51: 1481-1489..[CrossRef][Web of Science]
Frary A. Nesbitt T. C. Frary A. Grandillo S. van der Knaap E. Cong B. Liu J. Meller J. Elber R. Alpert K. B. Tanksley S. D.. 2000. fw2.2: a quantitative trait locus key to the evolution of tomato fruit size. Science 289: 85-88..
Freckleton R. P.. 2000. Phylogenetic tests of ecological and evolutionary hypotheses: checking for phylogenetic independence. Functional Ecology 14: 129-134..[CrossRef]
Galen C. Newport M. E. A.. 1987. Bumble bee behavior and selection on flower size in the sky pilot, Polemonium viscosum. Oecologia 74: 20-23..[CrossRef][Web of Science]
Garland T. Jr. Harvey P. H. Ives A. R.. 1992. Procedures for the analysis of comparative data using phylogenetically independent contrasts. Systematic Biology 41: 18-32..[Abstract]
Gillaspy G. Ben-David H. Gruisem W.. 1993. Fruits: a developmental perspective. Plant Cell 5: 1439-1451..
Grandillo S. Tanksley S. D.. 1996. QTL analysis of horticultural traits differentiating the cultivated tomato from the closely related species Lycopersicon pimpinellifolium. Theoretical and Applied Genetics 92: 935-951..[CrossRef][Web of Science]
Harder L. D. Barrett S. C. H.. 1996. Pollen dispersal and mating patterns in animal-pollinated plants. In D. G. Lloyd, S. C. H. Barrett, [eds.], Floral biology: studies on floral evolution in animal-pollinated plants, 140-190. Chapman and Hall, New York, New York, USA..
Harder L. D. Barrett S. C. H. Cole W. W.. 2000. The mating consequences of sexual segregation within inflorescences of flowering plants. Proceedings of the Royal Society of London, B, Biological Sciences 267: 315-320..
Harder L. D. Thomson J. D.. 1989. Evolutionary options for maximizing pollen dispersal of animal pollinated plants. American Naturalist 133: 323-344..[CrossRef][Web of Science]
Harvey P. H. Pagel M. D.. 1991. The comparative method in evolutionary biology. Oxford University Press, Oxford, UK..
Janzen D. H.. 1977. A note on optimal mate selection in plants. American Naturalist 111: 365-371..[CrossRef][Web of Science]
Joubés J. Phan T-H. Just D. Rothan C. Bergounioux C. Raymond P. Chevalier C.. 1999. Molecular and biochemical characterization of the involvement of cyclin-dependent kinase A during early development of tomato fruit. Plant Physiology 121: 867-869..
Knapp S.. 2002. Tobacco to tomatoes: a phylogenetic perspective on fruit diversity in the Solanaceae. Journal of Experimental Botany 53: 2001-2022..
Lee T. D.. 1988. Patterns of fruit and seed production. In J. Lovett-Doust, L. Lovett-Doust, [eds.], Plant reproductive ecology: patterns and strategies, 179-202. Oxford, University Press, New York, USA..
Levin R. A. Myers N. R. Bohs L.. 2006. Phylogenetic relationships among the "spiny solanums" (Solanum subgenus Leptostemonum). American Journal of Botany 93: 157-169..
Levin R. A. Watson K. Bohs L.. 2005. A four-gene study of evolutionary relationships in Solanum section Acanthophora. American Journal of Botany 92: 603-612..
Lloyd D. G.. 1979. Parental strategies in angiosperms. New Zealand Journal of Botany 17: 595-606..[Web of Science]
Lloyd D. G.. 1980. Sexual strategies in plants. I. An hypothesis of serial adjustment of maternal investment during one reproductive session. New Phytologist 86: 69-79..[CrossRef][Web of Science]
Lloyd D. G. Bawa K. S.. 1984. Modification of gender of seed plants in varying conditions. Evolutionary Biology 17: 255-338..[Web of Science]
Maddison W. P. Maddison D. R.. 2005. Mesquite: a modular system for evolutionary analysis, version 1.06. Website http://mesquiteproject.org [accessed 10 August, 2007]..
May P. G. Spears E. E. Jr. 1988. Andromonoecy and variation in phenotypic gender of Passiflora incarnata (Passifloraceae). American Journal of Botany 75: 1830-1841..[CrossRef][Web of Science]
Midford P. E. Garland T. Jr. Maddison W. P.. 2003. PDAP: PDTREE package for Mesquite, version 1.00. http://mesquiteproject.org/pdap_mesquite/ [accessed 10 August, 2007]..
Miller J. S. Diggle P. K.. 2003. Diversification of andromonoecy in Solanum section Lasiocarpa (Solanaceae): the roles of phenotypic plasticity and architecture. American Journal of Botany 90: 707-715..
Miller R. H.. 1969. A morphological study of Solanum mammosum and its mammiform fruit. Botanical Gazette 130: 230-237..
Nee M.. 1979. A revision of Solanum section Acanthophora. Ph.D. dissertation, University of Wisconsin, Madison, Wisconsin, USA..
Nesbitt T. C. Tanksley S. D.. 2001. fw2.2 directly affects the size of developing tomato fruit, with secondary effects on fruit number and photosynthate distribution. Plant Physiology 127: 575-583..
Podolsky R. D.. 1993. Evolution of a flower dimorphism: how effective is pollen dispersal by "male" flowers?. Ecology 74: 2255-2260..[CrossRef][Web of Science]
Posada D. Crandall K. A.. 1998. Modeltest: testing the model of DNA substitution. Bioinformatics 14: 817-818..
Price T.. 1997. Correlated evolution and independent contrasts. Philosophical Transactions of the Royal Society of London, B, Biological Sciences 352: 519-529..[CrossRef]
Primack R. B.. 1987. Relationships among flowers, fruits, and seeds. Annual Review of Ecology and Systematics 18: 409-430..[CrossRef][Web of Science]
Primack R. B. Lloyd D. G.. 1980. Andromonoecy in the New Zealand montane shrub manuka, Leptospermum scoparium (Myrtaceae). American Journal of Botany 67: 361-368..[CrossRef][Web of Science]
Purvis A. Rambaut A.. 1995. Comparative analysis by independent contrasts (CAIC): an Apple Macintosh application for analysing comparative data. Computer Applications in Biological Sciences 11: 247-251..
Rambaut A.. 2002. Se-Al: sequence alignment editor, version 2.0. Website http://evolve.zoo.ox.ac.uk/ [accessed 10 August, 2007]..
Ruiz Zapata T. Kalin Arroyo M. T.. 1978. Plant reproductive ecology of a secondary deciduous tropical forest in Venezuela. Biotropica 10: 221-230..[CrossRef][Web of Science]
SAS Institute.. 2002. JMP, version 5.0.1a. SAS Institute, Cary, North Carolina, USA..
Sokal R. R. Rohlf F. J.. 1995. Biometry, 3rd ed. Freeman, New York, New York, USA..
Solomon B. P.. 1985. Environmentally influenced changes in sex expression in an andromonoecious plant. Ecology 66: 1321-1332..[CrossRef][Web of Science]
Solomon B. P.. 1986. Sexual allocation and andromonoecy: resource investment in male and hermaphrodite flowers of Solanum carolinense (Solanaceae). American Journal of Botany 73: 1215-1221..[CrossRef][Web of Science]
Spalik K.. 1991. On the evolution of andromonoecy and ‘overproduction' of flowers: a resource allocation model. Biological Journal of the Linnean Society 42: 325-336..[CrossRef][Web of Science]
Spujt R. J.. 1994. A systematic treatment of fruit types. New York Botanic Garden, Bronx, New York..
Stanton M. L. Preston R. E. Ecological consequences and phenotypic correlates of petal size variation in wild radish, Raphanus sativus. American Journal of Botany 75: 528-539..
Stephenson A. G.. 1981. Flower and fruit abortion: proximate causes and ultimate functions. Annual Review of Ecology and Systematics 12: 253-279..[CrossRef][Web of Science]
Sutherland S.. 1986. Floral sex ratios, fruit-set, and resource allocation in plants. Ecology 67: 991-1001..[CrossRef][Web of Science]
Swofford D. L.. 2002. PAUP*: phylogenetic analysis using parsimony (* and other methods), version 4. Sinauer, Sunderland, Massachusetts, USA..
Tanksley S. D.. 2004. The genetic, developmental, and molecular bases of fruit size and shape variation in tomato. Plant Cell 16: S181-S189..
Vallejo-Marín M. Rausher M. D.. 2007. The role of male flowers in andromonoecious species: energetic costs and siring success in Solanum carolinense L. Evolution 61: 404-412..[CrossRef][Web of Science][Medline]
Veliath J. A. Ferguson A. C.. 1972. The effect of deblossoming on fruit size, yield, and earliness in tomato. Journal of Horticultural Science 7: 278-279..
Whalen M. D. Costich D. E.. 1986. Andromonoecy in Solanum. In W. G. D'Arcy, [ed.], Solanaceae: biology and systematics, 284-302. Columbia University Press, New York, New York, USA..
Whalen M. D. Costich D. E. Heiser C. B.. 1981. Taxonomy of Solanum section Lasiocarpa. Gentes Herbarium 12: 41-129..
Wyatt R.. 1982. Inflorescence architecture: how flower number, arrangement, and phenology affect pollination and fruit-set. American Journal of Botany 69: 585-594..[CrossRef][Web of Science]
Yampolsky E. Yampolsky H.. 1922. Distribution of sex forms in the phanerogamic flora. Bibliographia Genetica 3: 1-62..
CiteULike Complore Connotea Del.icio.us Digg Facebook Reddit Technorati Twitter What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | SEARCH RESULT |
