Heredity 85 (2000) 242±250 Received 19 November 1999, accepted 8 March 2000 F1 hybrid inviability in Eucalyptus: the case of E. ovata ´ E. globulus GUSTAVO A. LOPEZ à, BRADLEY M. POTTS* & PAUL A. TILYARD Cooperative Research Centre for Sustainable Production Forestry, School of Plant Science, University of Tasmania, GPO Box 252-55, Hobart, 7001, Tasmania, Australia and àInstituto Nacional de TecnologõÂa Agropecuaria (INTA) CIRN, Los Reseros y Las CabanÄas s/n 1712, Castelar (B.A.) Argentina The impact of inbreeding and hybridization on ®tness was compared in the two co-occurring forest tree species, Eucalyptus ovata and E. globulus, aimed at explaining the rarity of their hybrids in nature. The success of sel®ng, open-pollination and outcrossing of both species and interspeci®c hybridization was monitored from seed-set to 10-year's growth in a ®eld trial. There was a unilateral barrier to hybridization with seed-set obtained only with E. ovata females. The F1 hybrids exhibited reduced viability compared to intraspeci®c cross-types at virtually all stages of the life cycle and are clearly at a selective disadvantage compared with their open-pollinated E. ovata half-sibs with which they would directly compete in nature. Eucalyptus ovata and E. globulus overlap in their ¯owering time but the F1 hybrids ¯owered later with virtually no overlap with either species. The asynchronous ¯owering and reduced reproductive ®tness of F1 hybrids would markedly limit the opportunity for advanced generation hybridization. Inbreeding similarly had a deleterious eect on the ®tness of both species, and the F1 hybrids were most competitive with the E. ovata selfs. It is argued that changes in inbreeding levels of parental populations may be a key factor aecting the relative ®tness of hybrids and their potential to impact on the pure species gene pool. Reduced ®tness of the pure species through inbreeding may result in hybridization having its greatest evolutionary impact in small founder or relict populations. Keywords: Eucalyptus globulus, Eucalyptus ovata, hybridization, hybrid ®tness, inbreeding, reproductive isolation. Introduction of arti®cial hybrids with common environment trials is one solution, yet with tree taxa there is a paucity of Eucalypt species are well known for their weak repro- detailed, long-term studies. ductive barriers (Potts & Wiltshire, 1997). However, the The relative ®tness of hybrids and their parental taxa number of natural hybrid combinations recorded is may be habitat dependent or determined by endogenous relatively low given the opportunities for hybridization factors such as genomic incompatibilities which result in between sympatric taxa in nature (Grin et al., 1988). hybrids being less ®t than parent taxa in all environ- The extent of natural hybridization varies depending ments (Levin, 1978; Arnold, 1997; Emms & Arnold, upon numerous factors, including the degree of taxo- 1997). A key endogenous factor that has received little nomic and spatial separation, ¯owering synchrony, attention is the inbreeding level of parental species. This ¯ower size and hybrid ®tness (Grin et al., 1988; Gore is particularly important for plants with mixed mating et al., 1990; Ellis et al., 1991). Although there are many systems (e.g. Eucalyptus; Hardner & Potts, 1995) and in studies of natural hybridization in the genus (reviewed small populations, where natural interspeci®c hybrids in Grin et al., 1988; Potts & Wiltshire, 1997) only a few may compete with inbred pure species progenies. The address hybrid ®tness (e.g. Drake, 1981; Potts, 1986). present study examines pre- and postzygotic barriers However, these ®tness studies suer from a poor to hybridization between two species of Eucalyptus, knowledge of hybrid pedigree as identi®cation is mor- E. ovata Labill and E. globulus Labill ssp. globulus, phologically based (Arnold, 1992). Studying the ®tness aimed at explaining the rarity of their hybrids in nature. We report the success of sel®ng, open-pollination *Correspondence. E-mail: [email protected] and intra- and interspeci®c cross-pollination of both 242 Ó 2000 The Genetical Society of Great Britain. HYBRID INVIABILITY IN EUCALYPTUS 243 species and compare the relative eects of interspeci®c ted self-pollination), interspeci®c F1 crosses, as well as hybridization and inbreeding on ®tness from seed-set to open-pollinated (OP) controls were undertaken using 23 10-year's growth in a ®eld trial. (13 used as females for controlled crossing) E. globulus Eucalyptus globulus ssp. globulus (E. globulus here- and 12 (®ve were used as females for controlled crossing) after) and E. ovata are forest trees from dierent series E. ovata trees as parents. The numbers of crosses and (subgenus Symphyomyrtus; section Maidenaria; series families produced for each cross type are given in Viminales and Ovatae, respectively; Pryor & Johnson, Table 1. The pollen used for intraspeci®c outcrossing or 1971) which dier markedly in ¯ower morphology and interspeci®c F1 hybridization was derived from: (i) a size (Gore et al., 1990). Like most eucalypts, E. globulus mixture of pollen from ®ve E. globulus trees (polymix); has a mixed mating system, is pollinated by a variety of (ii) a mixture of pollen from ®ve E. ovata trees insect and bird taxa (Hingston & Potts, 1998) and (polymix); or (iii) single pollen collections. In each case, exhibits extreme inbreeding depression following sel®ng the pollen was collected from trees that were unrelated (Hardner & Potts, 1995). There is no information on the to the female parents. Crossing methodology, seed breeding system or eect of inbreeding on E. ovata. traits, germination and nursery procedures are detailed Natural hybridization between E. globulus and in Hardner & Potts (1995). E. ovata would be expected because they exhibit extensive overlap in geographical ranges and ¯owering Seed-set, germination and nursery growth time on the island of Tasmania (Williams & Potts, 1996). On a local scale they occupy dierent habitats but Cross success was assessed based on the number of grow in close proximity in ecotones where their crowns viable seeds per capsule, the number of viable seeds per may contact, yet hybridization is rare (Williams & Potts, pollinated ¯ower and seed viability (the ratio of viable 1996). Putative F1 hybrid seedlings have not been seeds to total seeds). Controlled germination tests were observed in open-pollinated progenies of E. globulus undertaken at 22°C with seeds that were classi®ed on but occur at a rate of »1% in open-pollinated progenies appearance as viable. Each seed lot (family) was of E. ovata sampled within »500 m of E. globulus partitioned into two separate Petri dishes containing a (unpubl. data). Natural hybrids in mature stands are maximum of 20 seeds which were allocated to two restricted to rare, isolated trees or small, localized randomized blocks within the germination chamber. patches of hybrids in ecotones (e.g. McAulay, 1937). Selfed seed from unassisted and assisted self-pollination treatments were pooled into a single seed lot at this stage. The proportion of seed germinated and the mean Materials and methods time to germination of each seed lot was calculated. After 25 days, healthy germinants were planted into Crossing design individual pots and maintained in family plots in a Three types of intraspeci®c controlled pollinations greenhouse. After 4 months, plants were transferred (assisted outcrossing, assisted self-pollination, unassis- outdoors for hardening and at age 6 months, just prior Table 1 Number of parents involved in the mating design, crosses carried Nursery Field trial out and number of families and Parents Crosses Families Seedling Families Seedling seedlings planted in the nursery and ®eld trial of Eucalyptus ovata, E. ovata E. globulus and their F1 hybrids Self 6 6 6 96 6 35 OP 5 5 12* 415 12 206 Outcross 9 14 14 647 14 216 F1 10 10 10 452 10 111 E. globulus Self 13 13 12 207 11 116 OP 20 20 20 656 20 284 Outcross 23 31 27 628 26 282 F1 18 28 0 0 0 0 Total 104 127 104 3101 99 1250 * Includes additional open-pollinated families from pollen parents. Ó The Genetical Society of Great Britain, Heredity, 85, 242±250. 244 G. A. LOPEZ ET AL. to setting up the ®eld trial, the number of runts (dwarfs compare: (i) the parental taxa under outcrossing; (ii) the and plants with extremely poor vigour), unhealthy and selfs and outcrosses within each taxon; (iii) the F1 hybrid healthy plants were assessed. Family arrangement within against the mid-parent value derived from the outcross the greenhouse and nursery was randomized with treatment for each parental taxon; and (iv) the F1 hybrids respect to cross type. against the open-pollinated progeny of E. ovata. The level of inbreeding depression (ID%) resulting from either sel®ng or open-pollination was calculated as: Field trial ID% 100(X ±X )/X A ®eld trial was established in 1988 near Ridgley in out inbred out north-western Tasmania (latitude 41°10¢S longitude where Xout is the mean for the controlled outcrosses and 145°46¢E) with healthy 7-month-old plants from the Xinbred is the mean for inbred progenies (self- or open- nursery. There were ®ve replicates, containing ®ve sub- pollinated). Outbreeding depression was measured in a blocks: (i) OP (GLop) and outcrosses (GL ´ GL) of similar manner as deviation below the average of the E. globulus; (ii) E. globulus selfs (GLself ); (iii) OP (OVop) E. ovata and E. globulus outcrosses. Where traits were and outcrosses (OV ´ OV) of E. ovata; (iv) E. ovata selfs transformed prior to analysis, the depression was (OVself ); and (v) E. ovata ´ globulus F1 hybrids calculated using back-transformed means. All analyses (OV ´ GL). This sub-blocking was chosen to minimize were undertaken with the GLM procedure of SAS (SAS competition eects between progenies from markedly version 6.12) and the Tukey test (P < 0.05) used for dierent cross types.