Fuchsia excorticata - a small tree from New Zealand
This species exhibits a floral-color change. Flowers start out green when they first open and change to red several days later. Pollinators visit the flowers when they are producing nectar - which is when they are green. They never visit the red flowers. We were drawn to this system because the pollinators are bellbirds (honeyeaters) and we had always learned that birds preferred to visit species with red flowers. Why then did this bird-pollinated species evolve green flowers?

Curt Lively and I found that this color change does not occur to draw the bird pollinators to the plant - pulling the red flowers off a tree does not affect visitation to the green flowers. Instead, the color change appears to have evolved as a cue, directing the birds away from the nectarless-red flowers and to the nectar-rich green flowers. Why is this cue necessary? The styles of this flower are quite long and it takes 3 days for the pollen tubes to grow down this style and into the ovary. This style falls off as a unit with the attractive structures of the flower, so the attractive structures must be retained during the period that the pollen is growing down the style. During this time the pollinators are directed away from the flower by its red color. See Delph and Lively 1989 Evolution 43:1252-1262.





Hebe strictissima - inflorescence from a female plant growing on Banks Peninsula (South Island of New Zealand). Females of this gynodioecious species occur in higher frequency in poor sites as a consequence of greater plasticity in fruit set on hermaphrodites (see Delph 1990 Evolution 44:134-142 and Delph 2003 Evolution & Development 5:34-39).


Sugarloaf Bush - home of Hebe strictissima










Silene acaulis - growing in the Arctic National Wildlife Refuge in Alaska. This gynodioecious species is circumpolar in the northern hemisphere. The plant pictured here is a hermaphrodite.

This species shows more known base-pair polymorphism in its mitochondrial genes than any other plant species in the world. We have hypothesized that this polymorphism is a consequence of balancing selection (see Städler and Delph 2002 PNAS 99:11730-11735). The gene that causes male sterility in this gynodioecious plant is located in the mitochondria and is likely to be under negative -frequency dependent selection. Since all genes in a mitochondria are supposed to be inherited as one large linkage group, variation in other genes is maintained as a consequence of selection on this male-sterility gene.



Silene latifolia - pistillate flower from a female on the left and staminate flower from a male on the right

I have been working on this species since my postdoc with Tom Meagher back in 1989. My main focus has been the factors promoting and constraining the evolution of sexual dimorphism, including stronger selection and more genetic integration in males than females, and within- and between-sex genetic correlations (many papers). I have also looked into the genomic architecture of sexually dimorphic traits (Scotti and Delph 2006 Evolution 60: 1793-1800; Delph et al. 2010 Evolution 64: 2873-2886). Currently, we are focusing on how local adaptation, sex-specific selection, sexual dimorphism affect sex-chromosome evolution with funds from the NSF.


Silene diclinis - an endangered dioecious species endemic to southeastern Spain. Mandy Brothers and I were able to produce hybrids between this species and Silene latifolia (see tag marking a cross). Silene diclinis is pollinated by bumblebees, whereas S. latifolia is pollinated by night-flying moths.
Mandy used F2s to determine the degree to which different floral traits impact pollinator visitation - including such traits as petal color (pink vs. white), petal size (small vs large), placement of flowers (inflorescence along the ground vs vertical inflorescence), and timing of flower opening (morning vs evening) (Brothers and Atwell 2014 International Journal of Plant Sciences).

We also found that Haldane's Rule holds for hybrids of these two species, in which males are the heterogametic sex (Brothers and Delph 2010 Evolution 64: in press), with dominance theory affecting male sterility (Demuth et al. 2014 Evolution).




Lobelia siphilitica - perfect flower from a hermaphrodite with the fused anther sac showing - pollen is contained within the sac and is extruded at the tip where a white frill of hairs can be seen.
This gynodioecious species has been extensively studied in the lab by me and two of my postdocs (Pia Mutikainen and Ben Montgomery) and two of my students (Dana Dudle and Maia Bailey).
Genetics of sex determination - sex in this species is nuclear-cytoplasmic, with multiple CMS (cytoplasmic male sterility) types and multiple restorers (see Dudle, Mutikainen, and Delph 2001 Heredity 86:265-276).
Inbreeding depression - although offspring from female lineages are likely to have less of a history of inbreeding, this does not result in greater inbreeding depression in these lineages. Rather, among-family variation overrides any variation between female and hermaphrodite lineages (see Mutikainen and Delph 1998 Evolution 52:1572-1582).
Cost of restoration - theory has shown that there must be a cost of restoration in order for nuclear-cytoplasmic gynodioecy to be maintained (see Bailey, Delph, and Lively 2003 Am Nat 161:762-776). Maia showed this cost was in terms of pollen viability in this species, and her work was the first to conclusively show a cost of restoration in a natural gynodioecious species (see Bailey 2002 Evolution 56:2178-2186).
Mitochondrial haplotype diversity - an investigation into how mt-haplotype diversity is related to the frequency of females within populations suggest that balancing selection maintains the CMS polymorphism needed for nuclear-cytoplamic gynodioecy in this species.


Gilia achilleifolia - this species exhibits within-population variation in the degree of herkogamy (separation of anthers and stigma), as shown here. This variation is associated with selfing rate - individuals with high herkogamy self less than individuals with low herkogamy.

This variation allowed us to test for an association between this floral trait and inbreeding depression. We found a positive association (Takebayashi and Delph 2000 Evolution 54:840-846), which is consistent with theory predicting that the history of inbreeding should affect inbreeding depression. In other words, the more inbred a lineage was in the past, the less inbreeding depression it should express.



Impatiens pallida - flower in its male phase.

Frank Frey, Ryan Davis and I have looked into
whether asymmetry in petal length might affect pollinator
visitation and hence the number of seeds produced
(note that the lower-left petal is longer than the
lower-right petal). We found that even drastic asymmetry
caused by experimental manipulation (cutting off 1/2 or
even all of the lower-left petal) did not affect seed
production in this bumblebee-pollinated species
(see Frey et al. 2005 International Journal of Plant
Sciences 166:659-662). This result constrasts with studies
suggesting that asymmetry will negatively affect plant fitness.
Note, however, that we only looked at seed fitness.









Claytonia virginica - This species, which is common in the woods around Bloomington, exhibits within-popultion polymorphism in flower color, ranging from pure white to dark pink.

My student Frank Frey studied how this polymorphism is maintained. Somewhat unexpectedly, he found that pollinators did not impose selection based on flower color. Flower color in this species is associated with two flavonols, quercitin and kaempferol, and white-flowered morphs have more of these compounds in their flowers and their leaves. Frank found evidence that opposing direction selection via herbivores and a fungal pathogen is a likely explanation for the maintenance of this polymorphism. See Frey 2004 Evolution 58:2426-2437, including cover photo.








Leavenworthia alabamica - This species is endemic to northern Alabama and exhibits among-population variation in its mating system. Some populations contain large-flowered, self-incompatible individuals (left) whereas others have small-flowered, highly self-fertilizing individuals (right). My advisor, David Lloyd, conducted his dissertation research on this and another closely related species to further our understanding of the ecological agents that favor self-fertilization in plants. One of my students, Jeremiah Busch, did his Ph.D. on this species.

There is a strong tendency for the self-fertilizing populations be small and inhabit ecologically marginal environments. Jeremiah found that the lack of mates in these environments favors the spread and fixation of self-compatible genotypes, which is followed by many adaptations for autonomous seed production (see Busch 2005 Am J Bot 92:1503-1512). He concludes that the floral transitions observed in the selfing populations of this species (small flowers, introrse anthers, short styles, and low pollen-to-ovule ratios) are driven by the benefits of reproductive assurance and the streamlining of adaptations for outcrossing. He also found evidence for heterosis in one of the small, isolated, selfing populations (Busch 2006 Evolution 60:184-191).


Leavenworthia torulosa (left)
and Leavenworthia stylosa (right) - both species are winter annuals endemic to limestone cedar glades in and around the Central Basin of Tennessee. Leavenworthia stylosa is found entirely within this geographic region and is self-incompatible, whereas L. torulosa may also be found in southern portions of Kentucky and is self-compatible. My student Ingrid Anderson studied floral development and selfing rates in these species.

She found that the two sister species exhibit suites of traits that typically differentiate selfing and outcrossing species: along with differences in flowering time, the outcrosser has larger flowers, while the selfer shows reduced flower size. Studies of correlation patterns suggest that pollinators select for increased floral integration in the outcrosser, and she also performed an artificial selection experiment on flowering time and selfing rate in L. torulosa.


Brassica napus - this crop species has been extensively studied regarding genes controlling cytoplasmic male sterility and their corresponding nuclear restorer genes. As a consequence, the molecular actions of the two well-characterized restorers have been characterized, providing a wonderful opportunity to directly test for the cost of restoration. We performed an experiment in which we examined individuals of known genotype at both the CMS loci and the restorer loci and found that the two restorer alleles in this species differ in the magnitude of their cost of restoration. This is the first direct test of a cost of restoration. Ben Montgomery, Maia Bailey and Lynda collaborated on this work. See Montgomery et al. 2014 Botany 92:847-853.


Thalictrum macrostylum - this species is cryptically dioecious: males look like males (their flowers contain stamens - photo on right) and females look like hermaphrodites (in addition to containing ovules, their flowers contain stamens with pollen that does not germinate - photo on left). Rebecca Penny, a former Ph.D. student in the lab, investigated why this enigmatic breeding system is maintained. She approached this question via field observations and experiments. See Penny 2014 IJPS 175:794-802.


DaturaDatura stramonium - Deidra Jacobsen, a Ph.D. student in the lab, used this species to investigate how growth, reproduction and defense were affected by herbivory and induced resistance in the absence of herbivory (via application of jasmonic acid).



Manduca sexta - Deidra Jacobsen has found that a high proportion of tobacco hornworms collected from nature are hosts to parasitoid wasps. She has established a colony from the field in the lab and is using it to compare to a old, established 'lab' colony in terms of how plant defensive compounds affect these herbivores.