Experimental Hybridization of Northern Chihuahuan Desert Region Opuntia (Cactaceae)

Possible natural hybridization amo ng II taxa of Opuntia sens u stricto was inve stig ated in the nonhero Chihuahuan Desert region through the use of experimental hybr idization. Established plant s representing specific taxa gro wing in the Sui Ross State Un iversity Opuntia garden were used for a ll experiment s . Reciprocal crosses were made between putati ve parental taxa of field-observed putative hybrids. and each experimental cross ana lyzed for fruit and seed set, For each taxon . test s were performed to control for possible apo mictic, autogamous. and ge itonogamous seed set. Several exper imental crosses were found to set seed in amounts expected for natural pollination events. Data gathe red from the tests also provided basic information regardin g the breeding systems of the taxa inve stig ated . Data presented here provide support for several hypoth esized hybridization events amo ng Opuntia. Key word s: Chihuahuan Desert , hybridization, Opuntia.


INTRODUCTION
Opuntia sensu lato is the largest genus within the subfamily Opuntioideae (Cactaceae), with an estimated 160 or more species (Gibson and Nobel 1986). Ninety-eight taxa of Opuntia within 44 species occur within the United States (Benson 1982). The most recent treatments reco gnize 24 species of Opuntia within the state of Texas with 22 taxa occurring in the Trans-Pecos (Powell 1998 ;Anderson 200 I) .
Three North American subgenera are traditionally recognized within Opuntia. The Cylindropuntia (chollas), Corynopuntia (club chollas), and Opuntia sensu stricto (prickly-pears) are distinguishable by habit and stem shape (Britton andRose 1919-1923;Benson 1982). Numerous authors have recommended that the subgenera of Opuntia should be elevated to generic rank (Robinson 1973 ;Anderson 1999, 200 I;Pinkava 1999 ). Representative species from each of these three subgenera are found within the Trans-Pecos. The prese nt study concerns itself only with plants of Opuntia sensu stricto. The prickly-pears form the bulk of the opuntioid taxa in the Trans-Pecos region, with 13 recognized species (Powell 1998) . Numerous prickly-pear populations, which are not easily accommodated within these 13 species, exi st within the Trans-Pecos region and adjacent areas. Hybridization between taxa is a common explanation for these populations.
Hybridization has been thought to give rise to new species, varieties, or morphotypes among Opuntia (Gibson and Nobel 1986 ;Anderson 200 I). Examples of putative hybridization abound in the literature. Opuntia kelvinensis has been de scribed as a clonal microspecies (G rant and Grant 1971 ) occurring in southern Arizona derived through hybridization between O. fulgida and 0. spinosior. Numerous examples of putative hybrid s have also been documented in Tran s-Pecos Texas. Tetraploid Opuntia X sp inosibacc a has been des cribed as a hybrid nothospecies derived from natural crossing of diploid O. aureispina and hexaploid O. phaeacantha (Pin kava and Parfitt 1988). Interploidal hybridization is believed to occur in Opuntia, resulting in even-ploid (Pinkava and Parfitt 1988 ) or odd-ploid (Grant and Grant 1982) progeny. Recent molecular work (Mayer et al. 2000) has elu cidated the hybrid origin of Opuntia Xp ro life ra, In addition to the above documented cases of putative hybridization, many additional workers have ob served and collected prickly-pears that exhibit intermediate morphology between described taxa, and were thought to be of hybrid origin. In previous studies, morphological , molecular, cytological and geographical data provided essentially all the evidence of hybridization . In spite of the mountain of references to natural hybrid Opuntia, I know of no do cumentation that transfer of pollen between taxa can result in seed set. Although strong evidence exists for the hybrid origin of certain cacti through the use of artifici al hybridizations (Powell et al. 1991;Powell 1995;Powell 1999), putative hybrid Opuntia have not been documented through artificial hybridization experiments.
Several specific cases of possible natural hybridiza-tion in Trans-Pecos Opuntia were identified and brought under investigation during the current study: Opuntioid floral morphology early in the flowering season promotes natural outcrossing (Grant J979). Early flowers were selected to use for the arti ficial crosses to reduce the possibility of self-pollination. Flower buds were bagged before anthesis with a 900 ern? (30 cm X 30 ern) piece of double-ply cheesecloth to deter pollination by floral visitors. Bagged flowers were checked for anthesis twice daily, at 1000 hrs and 1500 hrs. Appropriate pollen was transferred to open flowers at these times. Pollen transfer was performed with disposable cotton swabs that were broken in half after use to prevent cross-contamination. A set of three controls was designed to check for interfertility. To control for autogamy, flowers were bagged and pollen was transferred from the donor flower to its own gynoecium. To control for geitonogamy, pollen was transferred from several flowers of the plant to stigmas in different flowers on the same plant. To control for apomictic seed set, flowers were bagged, and no pollen transfer was performed. During spring 2000 the flowers of certain species were emasculated. The objectives of the emasculation procedure were to test for apomixis, expected to occur in certain hexaploids such as O. engelmannii. Flowers were emasculated prior to anthesis and prior to anther dehiscence, and bagged with cheesecloth immediately following the procedure. Emasculation was carried out by circumcision of the tepals above the attachment point to the pericarpel, followed by careful removal of the androecium with small forceps. Collecting fruits derived from unbagged flowers at the end of the bloom period obtained data relev ant to the ex pec ted seed set for each indiv idual. Th ese data were used for comparison with the arti ficial crosses a nd w ith the co ntrols. Fruits that dev eloped fro m open poll inat ed fruits are her eafter referred to as " native fru its ," Fruit and seed se t evaluation.-Floral product s were collec ted as they matured, and evaluated visu all y for fruit se t. Fruit set was scored as either positive or negati ve . Seeds were dis sected out of the fruits and counted. Mean and standard deviation of seed number were calculated for each test.
Analysis of hybrid embryos.-A ny live embryos derived from experimental crosses were determined to be hybrid embryos if two co nditio ns were met: ( 1) If normal seed se t was obser ved in native fruit s o f the same pl ant ; and (2) if virtu ally no seed set was observed in the apo m ictic, autogamou s, and geit onogamous tests on the sa me plant.

RESULTS
Fruit and seed set sco ring .-N ative fruit set occurred as e xpect ed in all 13 taxa of Opuntia in the current study ( O. a ureisp ina and O. Xs p inosibacca, and between O. chisosensis and O. Xs pi nos ibacca. Healthy mature fruits were always pres ent whe n seed set numbers were above zero. Conversely , all floral products that did not develop into healthy fruit s inv ari ably lacked seed s. Seed counts for all tests performed are listed in Table 2.     specific test. Fully developed fruits always contained numbers of viable seeds that were consistent with expected seed numbers in natural populations. Shriveled pericarpels, i.e., undeveloped fruits , invariably did not contain any seeds. The experimentally demonstrated interfertility between O. aureispina and O. macrocentra "azurea type" supports the hypothesis (Griffith 2001) that O. Xrooneyi is the result of hybridization between these parental taxa.
Results from the artificial crosses between O. engelmannii var. engelmannii (2n = 66) and O. engelmannii var. lindheimeri (2n = 66) are inconclusive regarding the interfertility of these two taxa. Both taxa exhibited abundant seed set for every test performed . No test involving var. engelmannii or var. lindheimeri resulted in complete prevention of seed set. Apomixis trials show that these taxa do set seed in large numbers through apomixis ( Table 2). The seeds resulting from artificial reciprocal crosses between the two varieties are likely the result of apomixis as well.
I expect the partially sympatric taxa O. atrispina and O. strigil to be fully interfertile in the field. The experimentally demonstrated interfertility supports the hypothesis that plants observed in areas of sympatry exhibiting intermediate morphology between O. atrispina and O. strigil are hybrids between these two taxa. Interfertility may also support the hypothesis that 0. atrispina and 0. strigil represent two ends of a morphological cline within one variable species.
Opuntia aureispina and O. chisosensis appear to be fully interfertile. The only barrier to the hybridization of O. aureispina and O. chisosensis in the field appears to be the distance between the populations. The experimentally demonstrated interfertility suggests a close relationship between these two taxa.
The documented interfertility (Table 2) between O. macrocentra "azurea type" and O. rufida suggests an explanation for a natural population of plants exhibiting intermediate morphology between these taxa. These results support the hypothesis that the natural intermediate population could result from interspecific hybridization of these taxa.
Seed set data suggest that artificial hybridization was successful and that natural hybridization is possible among members of the Opuntia macrocentra complex. The artificial crosses between O. macrocentra var. macrocentra (2n = 22) and O. macrocentra "azurea type " (2n = 22) document the interfertility of these plants (Table 2). I expect O. macrocentra var. macrocentra and O. macrocentra "azurea type" to be fully interfertile under natural sympatric conditions. Complete interfertility was not observed in crosses between O. macrocentra var. minor (2n = 44) and O. macrocentra var. macrocentra, or between O. macrocentra var. minor and O. macrocentra "azurea type." Experimental crosses using O. macrocentra var. minor as the female parent were not found to set seed. The seed set data for O. macrocentra var. minor, O. macrocentra var. macrocentra, and O. macrocentra "azurea type" suggest a directional barrier to hybridization related to ploidy level. Previous tests (Lewis 1979) demonstrated a similar directional barrier to reproduction among Pyrus (Rosaceae: Amygdaloideae) specimens of different ploidy levels. As O. macrocentra var. minor is sympatric or peripatric with both O. macrocentra var. macrocentra and O. macrocentra "azurea type," it is possible that hybrid triploid plants exist in the field . These data support previously reported odd-ploid chromosome counts (Grant and Grant 1982).
Crosses between Opuntia engelmannii var. engelmannii (staminate parent) and O. X sp inosibacca (ovulate parent) resulted in abundant seed set (