Forest Science, Vol. 34, No. I, pp. 139-151. Copyright 1988 by the Society of American Foresters Hybridization of the California Firs

William B. Critchfield

Abstract. Four groups of firs (sections, in the most recent classification of Abies) are represented in California. Crossing within these sections is possible and even easy, and in two of the sections intergrading populations between highly crossable taxa are wide spread in California. An exception is A. amabilis, a Northwestern fir that has not been crossed with other species in the same section {Grandes: A. concolor, A. grandis) or in other sections (e.g., Nobiles: A magnified). Crossing species in different sections is usually difficult or impossible. The genetic isolation of A. bracteata, an endemic species classified as a monotypic subgenus or section, may be nearly complete: two probable hybrids with A. concolor died at a few years of age. A few putative hybrids from inter- sectional crosses between species in Grandes and Nobiles died within months of germi nation. Intersectional crosses with firs outside California (two Mexican and four Eur asian species) all failed except A. concolor x A. religiosa, which produced numerous healthy hybrids. The common occurrence of genetic barriers in Abies is at odds with the long-held view that it is easy to hybridize fir species. For. Sci. 34(1): 139-151. Additional key words. Abies, interspecific hybrids, crossability, classification.

The ability of species to hybridize has not been explored as systemati cally in the genus Abies (true firs) as it has in other genera of Pinaceae such as Pinus and Pice a. A flurry of interest in Abies hybridization more than two decades ago (Klaehn and Winieski 1962, Mergen et al. 1964, Rohmeder 1961) had two results: it generated long lists of ostensibly successful hy brids, including many between Eastern and Western Hemisphere species, and it reinforced the impression that "there is little difficulty in hybridizing fir species" (Mergen et al. 1964). These papers provided little data on species crossability, however, and nearly all of the hybrids claimed as suc cesses were unverified (Hawley and DeHayes 1985a). Mergen et al. (1964), for example, counted 18 species combinations as successful new crosses (four between Eurasian and North American species, one with both North American parents, and the rest with both Eurasian parents), but none of the putatively hybrid seedlings were compared with their parent species. The first full account of crossing behavior in a taxonomic group of firs was that of Hawley and DeHayes (1985a). Three North American species constitute section Balsameae in the classification of Liu (1971): subalpine fir (A. lasiocarpa [Hook.] Nutt.), balsam fir (A. balsamea [L.] Mill.), and Fraser fir (A. fraseri [Pursh] Poir.) Hawley and DeHayes found that these taxa were moderately to fully crossable with each other, but most crosses with white fir (A. concolor Hildebr.), a North American species in another section, either failed or produced small numbers of hybrids. This paper summarizes the results ofa program of exploratory hybridiza tion in Abies carried out between 1962 and 1974 at the Forest Service's Institute ofForest Genetics (IFG), located in the Sierra Nevada foothills

The author is geneticist at the USDA Forest Service Pacific Southwest Forest and Range Experiment Station, Box 245, Berkeley CA 94701. B. Burr, L. Y. Hsin, D. I. Saybold, and J. B. St. Clair studied and described fir hybridswhile they were summer interns at the Institute of Forest Genetics, Placerville, CA. R. T. Bingham, J. F. Franklin, E.L. Little, Jr., and R. J. Steinhoff generously provided pollen of native firs from the Pacific Northwest, Japan, and Mexico. Manuscript received July 21, 1987.

March 1988/ 139

Reprinted from the Forest Science, Vol. 34, No. 1, March 1988 near Placerville, California. Six firs native to California were one or both parents of all crosses. Most crosses were made on red fir (A. magnifica A. Murr.) or white fir in natural stands. Pollen parents included, in addition to the California firs, six species native to Mexico, Spain, or eastern Asia. Among the crosses made at IFG were the first attempts to hybridize the unique bristlecone fir (A. bracteata Poiteau), a California endemic. Our data corroborate the recent findings of Hawley and DeHayes (1985a) and Kor- mutak (1981) that genetically controlled interspecific barriers expressed during sexual reproduction are present in Abies, but the data are still too limited to establish whether these barriers are as important in this genus as they are in other genera of Pinaceae.

DISTRIBUTION AND CLASSIFICATION

Two of the six firs native to California are present only as small outliers in the Klamath Mountains of northern California (Griffin and Critchfield 1976): subalpine fir, which is wide-ranging in western North America outside Cali fornia, and Pacific silver fir (A. amabilis Forbes) of the Pacific Northwest. Two others, white fir and grand fir (A. grandis [D. Don] Lindl.) constitute a species-complex that is widely distributed in the western United States and well represented in California. Grand fir, a primarily Northwestern species, extends south along the California coast to near the mouth of the Russian River. White fir is distributed from California to the southern Rocky Moun tains. California white fir, sometimes designated A. concolor var. lowiana (Gord.) Lemm. to distinguish it from Rocky Mountain white fir (var. con- color), has its principal distribution at middle elevations in the Sierra Ne vada. A geographic variant of uncertain status is present in the mountains of southern California. In the mountains of northwestern California (and northeast to Idaho) are populations that intergrade between grand and white firs. In eastern California, a few western outliers of Rocky Mountain white fir grow in isolated mountain ranges of the Mojave Desert. Red fir is also part of a species-complex. In its typical form it is restricted to upper eleva tions in the Sierra Nevada, and at the lower edge of its elevational range it overlaps white fir. In northern and northwestern California, red fir inter- grades with noble fir (A. procera Rehd.) of the Pacific Northwest, and these variable populations are sometimes called Shasta red fir (A. magnifica var. shastensis Lemm.). Bristlecone fir has a limited range in the coastal moun tains of central California, where it grows at lower elevations than most other firs. The diversity of the six firs native to California is recognized in all three recent monographs of Abies (Franco 1950, Gaussen 1964, Liu 1971). They are placed in 4 of 12 groups by Franco and in 4 of 15 groups by both Gaussen and Liu (Table 1). Bristlecone fir is the only species in one of two subgenera in the classifications of Franco and Liu, and the only species in one of five unnamed sections in Gaussen's less formal classification. Subal pine fir and its eastern relatives are a separate group in all three classifica tions. The red fir-noble fir complex is also grouped separately by Liu and Franco, but Gaussen surprisingly separates red fir from a group comprising Shasta and noble firs (Table 1). Pacific silver fir is grouped with white and grand firs in all classifications. The 12 species used in the crosses at IFG provide a fairly good sample of Abies. Among Franco's (1950) 12 groups, 9 are represented, and among the 15 groups of Gaussen (1964) and Liu (1971), 9 and 10 are represented. Each

140/ Forest Science TABLE 1. Recent classifications of Abies, including species names mentioned in the text"

Franco 1950 Liu 1971 Gaussen 1964

Subgen. Pseudotorreya (NA) Subgen. Pseudotorreya (NA) Sect. I + bracteata Sect. Bracteatae group 1 (NA): balsamea, Subgen. Sapinus + bracteata fraseri, + lasiocarpa Sect. Nobiles (NA) Subgen. Abies group 2 (A): sibirica, other + magnified, procera Sect. Nobiles (NA) group 3 (A): koreana, Sect. Oiamel (NA) + magnifica, procera sachalinensis, veitchii, religiosa, vejari, other Sect. Oyamel (NA) other Sect. Balsameae (NA) religiosa, other group 4 (A): homolepis, Ser. Grandes Sect. Vejarianae (NA) other + amabilis, vejari group 5 (A): firma + concolor, Sect. Grandes (NA) Sect. II (E) + grandis, other + amabilis, + concolor, group 1: cilicica, pinsapo, Ser. Lasiocarpae + grandis, other other balsamea, fraseri, Sect. Balsameae (NA) group 2: alba, + lasiocarpa balsamea, fraseri, cephalonica, Sect. Pichta (A): koreana, + lasiocarpa nordmanniana, other sachalinensis, sibirica, Sect. Pichta (A): sibirica Sect. Ill (A) veitchii, other Sect. Elate (A): koreana, group 3: forrestii, other Sect. Momi (A) veitchii, sachalinensis, group 4: mariesii, other Ser. Firmae: firma other Two other groups Ser. Homolepides: Sect. Momi (A): firma Sect. IV (NA) homolepis, marie sii, Sect. Homolepides (A) group 1: + amabilis, other homolepis, marie sii, + concolor, +grandis, Ser. Sinenses:/orr^5//7, other + magnifica, other other Sect. Elateopsis (A) group 2: religiosa, vejari, Sect. Peuce (E) forrestii, other other Ser. Albae: alba, Sect. Abies (E): alba, group 3: procera, cephalonica, cephalonica, + shastensis nordmanniana, other nordmanniana, other Sect. V (NA): + bracteata Ser. Pinsapones: Sect. Piceaster (E) cilicica, pinsapo, cilicica, pinsapo, other other Two other sections (A) One other section (A)

a NA = North America, E = Europe, A = Asia, + = native to California classification is based on a series of hypotheses about relationships of the firs, and our crosses provide a preliminary test of some of these hypotheses.

REPORTS OF NATURAL AND ARTIFICIAL HYBRIDIZATION

Several species-complexes in Abies include intergrading populations that are usually interpreted as having arisen through secondary intergradation: hybridization and introgression of formerly allopatric taxa. According to the alternative explanation—primary intergradation—the same taxa have always been in contact and are as distinct now as they have ever been. Most evidence supports the hypothesis of secondary intergradation for the white fir-grand fir complex (Hamrick and Libby 1972, Zavarin et al. 1977) and for the intermediate populations of subalpine and balsam firs in Alberta (re viewed in Critchfield 1984b). It is less certain that the red fir-noble fir com plex has evolved in the same way; the fragmentary fossil record of the group

March 1988/ 141 provides some support for the hypothesis of primary intergradation (Zavarin et al. 1978). The other two California firs are not known to hybridize in nature. The range of bristlecone fir does not overlap that of any other fir, but Pacific silver fir is sympatric with the other firs of the Pacific Northwest. Nearly a century ago dendrologist C. S. Sargent found a single tree that he consid ered a hybrid between Pacific silver and subalpine firs in the Olympic Mountains of Washington, but the specimen he collected does not support his identification of the tree as a hybrid (Critchfield 1977). The artificial hybrid between California white fir and grand fir (Table 2), produced in 1924 by controlled pollination on an arboretum tree in Den mark, was the first artificial hybrid of a California fir and among the first in

TABLE 2. Reported artificial hybrids of California species of Abies.

Female parent Male parent Reference

VERIFIED:* balsamea concolor Hawley & DeHayes 1985b balsamea lasiocarpa Hawley & DeHayes 1985b balsamea var. concolor Hawley & DeHayes 1985b phanerolepis Fern. balsamea var. lasiocarpa Hawley & DeHayes 1985b phanerolepis concolor var. concolor religiosa St. Clair & Critchfield in prep. concolor var. lowiana grandis Larsen 1934 fraseri concolor Hawley & DeHayes 1985b fraseri lasiocarpa Hawley & DeHayes 1985b

magnifica procera Silen et al. 1965

procera magnifica Silen et al. 1965 UNCERTAIN: cephalonica concolor Kantor&Chira 1971 cephalonica grandis Kantor&Chira 1971 concolor nordmanniana Rohmeder 1961

concolor procera Rohmeder 1961 concolor veitchii Rohmeder 1961 UNVERIFIED: alba concolor Rohmeder 1961 alba grandis Rohmeder 1961 concolor alba Rohmeder 1961 concolor amabilis Rohmeder 1961 concolor sibirica Ledeb. Rohmeder 1961 grandis alba Rohmeder 1961 grandis nordmanniana Rohmeder 1961 grandis veitchii Rohmeder 1961 homolepis Sieb. & Zucc. concolor Rohmeder 1961 homolepis grandis Rohmeder 1961 koreana lasiocarpa Mergen et al. 1964 nordmanniana concolor Rohmeder 1961 nordmanniana grandis Rohmeder 1961

procera concolor Rohmeder 1961

procera lasiocarpa Mergen et al. 1964 sachalinensis lasiocarpa Mergen et al. 1964 veitchii concolor Rohmeder 1961 veitchii grandis Rohmeder 1961

a Verified: Hybridity convincingly established. Uncertain: Description of putative hybrid sketchy and incomplete. Unverified: Putative hybrid listed but not described.

142/ Forest Science the genus. Larsen (1934) obtained nine trees from seeds of a single cone, and identified them as hybrids by "the bright green and shining needles so characteristic of the Father tree, the A. grandis." Other verified hybrids of white fir were produced more recently: Rocky Mountain white fir x oyamel (A. religiosa [H.B.K.] Schlecht. & Cham.) at the IFG in 1967 (St. Clair and Critchfield in prep.), and, in Vermont during the years 1980 through 1982, hybrids between balsam and Fraser fir females and white fir males of unknown origin (Hawley and DeHayes 1985a,b). In the same Vermont program, subalpine fir was the pollen parent in successful crosses with balsam and Fraser firs (Table 2). Hybrids of red and noble firs, produced in Oregon and at the IFG from 1962 crosses in both directions, were identifiable soon after germination by their intermediate cotyledon number (Silen et al. 1965). Most of the putative hybrids produced by Rohmeder (1961) between white or grand firs and other species were undescribed and undocumented except for the year or years in which the crosses were made (Rohmeder's Figure 14). They are listed in Table 2 as Unverified. Three exceptions, all made on white fir females, are listed in Table 2 as Uncertain. They are placed in this category because their documentation is incomplete; it in cludes little quantitative data and few details of any kind. In these three species combinations Rohmeder distinguished two kinds of hybrids: (a) "concolor-like" and (b) "green-needled." In all three hybrid comparisons, the white fir controls were taller than the pollen-parent controls at nine years. In each comparison, the group (a) trees were similar in height to white fir, and can probably be attributed to pollen contamination. The height of the three (b) groups, however, exceeded that of the white fir con trols by 72-111%. Limited morphological data supported the hybrid identi fication of the combinations with pollen parents A. nordmanniana Spach. (needle shape, bud resin) and A. veitchii Lindl. (needle waxiness). The identity of the white fir x noble fir hybrids is more questionable, although the "green-needled" group was more than twice as tall as controls of either parent species (Rohmeder 1961, Figure 17). White and noble firs have abun dant stomata on both leaf surfaces, and it is unlikely that their hypothetical hybrid offspring would fit the description of "green-needled." Two other combinations listed as Uncertain in Table 2 are putative hy brids of Greek fir (A. cephalonica Loud.) with white and grand firs. Both hybrids were produced in small numbers from crosses accompanied by suc cessful within-species control crosses. Hybrid and nonhybrid frequency distributions of first-year height only partly overlapped, and the hybrids averaged 9-30% taller than their Greek fir half-sibs (Kantor and Chira 1971). The list of Unverified hybrids in Table 2 also includes three combinations with subalpine fir as pollen parent (Mergen et al. 1964). The noble fir x subalpine fir cross produced only three seedlings, but the other two crosses, both on Asian species (A. koreana Wilson, A. sachalinensis Masters), yielded moderate to large numbers of putatively hybrid seedlings per cone harvested. Several unsuccessful crosses between species of Abies have been re ported in sufficient detail to establish that genetically controlled break downs can occur during reproduction (Table 3). Some of the failed crosses of white fir, including white fir x balsam and Fraser firs, were successful with white fir as pollen parent (Table 2). Other failed crosses duplicate com binations for which unverified hybrids have been reported (Table 2: A. con-

March 1988/ 143 TABLE 3. Failed interspecific crosses involving California firs.

Female parent Male parent Reference Comments

concolor balsamea Hawley & DeHayes 1985a Successful control crosses on both female parents concolor balsamea var. Hawley & DeHayes 1985a Same as above phanerolepis concolor fraseri Hawley & DeHayes 1985a Same as above concolor lasiocarpa Hawley & DeHayes 1985a Same as above concolor var. alba Kormutak & Dubovsk^ 1984 Repeated attempts; lowiana postfertilization block: proembryo abortion grandis alba Kantor & Chira 1971 Two of three pollen parents in native stand; successful control cross on female parent grandis cilicica Cam Kantor & Chira 1971 Successful control cross on female parent grandis nordmanniana Kantor & Chira 1971 Same as above grandis pinsapo Boiss. Kantor & Chira 1971 Same as above pinsapo concolor var. Kormutak 1981, Kormutak Prefertilization block: lowiana & Dubovsky 1984 inhibition of pollen tube growth

color and A. alba Mill., A. grandis and A. alba, and A. grandis and A. nordmanniana). Reproductive barriers expressed both before and after fer tilization have been found between California white fir and two European firs, A. pinsapo and A. alba (Table 3).

MATERIALS AND METHODS

Most of the crosses reported here were made in central Sierra Nevada forests, but a few maternal parents were arboretum trees at IFG (Table 4). Techniques worked out for pines in the 1930s (Cumming and Righter 1948) were used for controlled pollination of the firs. Pollen was collected in nat ural stands and in arboreta in San Francisco, Sacramento, and at IFG. Peak pollen shedding of white fir is usually one to two weeks earlier than in red fir, and in one season pollination of white fir in a mixed stand at 2300-2400 m was 13 days earlier than in red fir. Unless otherwise noted, crosses in Tables 5 through 7 were made with fresh or one-year-old deep-frozen pollen, and were accompanied by suc cessful control crosses. Control crosses were made on the same maternal parent in the same season as crosses between taxa; the pollen parent was the same taxon or race as the maternal parent; and germinable seeds were produced. We were unsuccessful in separating sound and hollow fir seed, and for most crosses all seeds harvested were planted in the nursery. The firs, unlike the pines, have cones and seed coats that develop without pollination. Wright (1953) obtained 13 cones with a nearly full complement of full-sized seeds from 20 bagged but unpollinated female strobili of an un specified fir. We bagged but did not pollinate 30 female strobili on three native white firs, and 21 (70%) developed into mature cones. Control crosses on the same trees yielded 23 cones (77%) from 30 strobili pollinated with white fir pollen. Thus ovule and conelet abortion do not provide useful indices of crossability in the firs as they do in the pines, and in this paper the

144/ Forest Science TABLE 4. Parent trees used in controlled crosses of California firs.

Number of parents

Species Female Male Location or geographic origina

Natural stands: amabilis 2 Oregon Cascades bracteata 1 Central California concolor var. lowiana 8 7 Central Sierra Nevada (5) grandis 6 N Idaho (2) grandis 8 W Oregon grandis 2 Oregon Cascades grandis 3 N California coast lasiocarpa 4b Oregon Cascades (2) magnifica 4 4 Central Sierra Nevada religiosa 1 Central Mexico veitchii 1 Japan: Mt. Fuji vejari Mart. 1 NE Mexico Arboretum trees: bracteata 1 5 Unknown concolor var. concolor 3 2 Colorado concolor var. lowiana 2 Sierra Nevada firma Sieb. & Zucc. 1 Unknown forrestii C.C. Rogers 2 China forrestii 1 Unknown magnifica 1 Unknown pinsapo 1 Unknown

a Number of localities in parentheses. b Minimum number in pollen mixes.

sole measure of crossability is the mean yield of germinable seeds per cone, expressed as a percentage of the yield of germinable seeds per cone in con trol crosses. The effect of pollen age was not part of this study, but repeated use of the same pollen and simultaneous use of pollen of different ages in white fir x grand fir crosses showed that deep-frozen fir pollen retains part or all of its ability to produce viable seed for at least three years. In crosses on four white firs, grand fir pollen produced more germinable hybrid seeds per cone after two years in a deep freezer than in the year it was collected, and deep- frozen pollen from another grand fir stand produced more seed after three years than it had after one year. In another set of three paired crosses, how ever, three-year-old pollen produced less than half as much germinable seed as one-year-old pollen from a different source.

RESULTS

Crosses Among the California Firs

Within-species crosses yielded fewer germinable seeds per cone in red fir than in white fir: 43.9 and 56.4 (Table 5). Red fir averaged much larger total numbers of seeds per cone, however (341 vs. 145), and a much lower percentage of germinable seeds—13% in red fir compared with 39% in white fir. Red fir is isolated from the white fir-grand fir group by nearly complete barriers expressed during sexual reproduction. Crosses in one direction, white fir x red fir, produced no offspring (Table 5). When red fir was the

March 1988/ 145 TABLE 5. Crosses among California firs.

Male parent

Female parent White fir Grand fir Red fir Bristlecone fir

White fir 4, 4 (2) 6a See Table 6 5, 4 (2) 7 7, 6 (3) 9b 36, 56.4 42,0 55, 0.04 Red fir 5, 7 (2) 17 3, 3 (2) 5C 5, 4 (2) 7 3, 3 (2) 3 129, 0.03 32, 0.2 38, 43.9 18,0 Bristlecone fir 1, 1 (1) 1 self (1) 1 1,0 1,92

a Number of: female parents, male parents (years) total crosses

cones harvested, mean germinable seeds per cone b Three crosses on trees of Rocky Mountain origin. c Three crosses with northern Idaho pollen. maternal parent, however, eight seedlings were obtained from three crosses with white or grand firs, all in the same year (1962). The seedlings began to die soon after germination, and all were dead by the end of the summer. This early mortality may have been due partly to the harsh environment of nearly empty nursery beds. In two progenies, the number and length of cotyledons were within the range of red fir, and there was no evidence that the seedlings were hybrids. The third progeny, two seedlings from a cross between red fir and a pollen mix of three Idaho grand firs, had 6 and 7 cotyledons. The seedlings were dwarfed, and their cotyledons were shorter than those of either parent species (Franklin 1961). Red fir seedlings from the same stand as the maternal parent averaged 8.8 cotyledons in a 42-seed- ling sample that included only two seedlings with 7 cotyledons and none with 6 (Silen et al. 1965). Grand fir has fewer cotyledons than red fir, usually 5 or 6 but ranging from 4 to 7 (Franklin 1961). The diminutive size and particularly the intermediate cotyledon number of these two short-lived seedlings suggest that they were true red x grand fir hybrids. Crosses of red and white fir females with bristlecone fir pollen parents were unsuccessful with one exception (Table 5). A 1967 Rocky Mountain white fir x bristlecone fir cross produced two seedlings recognized as prob able hybrids soon after germination. Both seedlings had 6 cotyledons, com pared with 6-9 in their white fir half-sibs. The needles of the hybrid seed lings were significantly shorter than those of the white fir seedlings (means 14.0 and 18.2 mm) and, unlike the white firs, had no stomata on the upper (adaxial) surface. The hybrids lived for 5 and 7 years. The longer-lived tree was only 24 cm tall at 5 years, with smaller leaves than those of either parent species. A single reciprocal cross, bristlecone fir x white fir, also failed to produce germinable seed (Table 5). The accompanying "control," made with self- pollen, yielded 92 germinable seeds from one cone. This high self-fertility, if it is not a chance occurrence, may be related to the unusually low genetic variability of bristlecone fir (Ledig 1987). Both Sierra Nevada and Rocky Mountain races of white fir were used as female parents in crosses with grand fir and each other. Grand fir pollen parents represented several widely separated geographic sources (Table 4), but the possible influence of geographic origin on crossability was obscured by differences in pollen age. Overall, crosses of grand fir with the races of white fir averaged about 86% as many germinable seeds as the control crosses (Table 6). Crosses between the races of white fir were less successful than the white

146/ Forest Science TABLE 6. Crosses in the white fir-grand fir group.

Male parent

White fir

Female parent Sierra Nevada Rocky Mts. grand fir

White fir Rocky Mts. 3, 2(l)5a 2, 1 (1)2 3,3(1)5 49, 28.4 21,38.8 63, 35.8 Sierra Nevada 4, 4 (2) 6 3, 2 (2) 5 4, 16 (2) 24 36,56,4 29, 26.4 163, 45.6

a Number of: female parents, male parents (years) total crosses

cones harvested, mean germinable seeds per cone fir x grand fir crosses, but fewer parents were involved, and the Rocky Mountain trees were growing outside their native range. Crosses of Cali fornia and Rocky Mountain trees in both directions produced fewer germin able seeds per cone than did intraracial control crosses (Table 6). Rocky Mountain x Sierra Nevada crosses yielded only 73% as much seed as within-Rocky Mountain crosses, and the reciprocal combination produced only 47% as much germinable seed as Sierra x Sierra crosses. The white x grand fir hybrids produced at IFG were easily identified by the same feature Larsen (1934) used to identify the first hybrids of this com bination. At 19 years the remaining IFG hybrids, all from maternal parents of Rocky Mountain origin, had needles with shiny green upper surfaces nearly devoid of stomata. The few stomata present were concentrated near the tip (which was usually notched, like that of grand fir), or in a single broken or continuous line extending a few mm from the tip, or—rarely—in a line extending to the needle base. A few crosses of white and red firs with pollen of subalpine and Pacific silver firs from the Cascade Mountains of Oregon failed to yield any germin able seeds (Table 7).

Crossing the California Firs With Other Species

Our attempts to cross red, white, and grand firs with Eurasian firs were uniformly unsuccessful, producing no germinable seeds, and all but one of the crosses with Mexican firs also failed (Table 7). These attempts were on a small scale, with one to three crosses for each species combination. Some of the pollens had been frozen for three years or more (four in the case of A. vejari), and most of the pollens were not tested in within-species crosses. Collectively, however, the crosses listed in Table 7 represent a fairly sub stantial demonstration that not all firs are easy to hybridize. The single success was the combination of white fir and the Mexican oyamel (A. religiosa). Like other hybrids between white fir and species lacking stomata on the upper surface of their leaves, the white fir x oyamel hybrids were most readily identified by their similarity to the pollen parent in this feature (St. Clair and Critchfield, in prep.).

DISCUSSION

What has been learned in recent years about crossing behavior in Abies negates the view that firs in general hybridize freely and provides some support for the geographically based classifications of Franco (1950) and Liu

March 1988/ 147 TABLE 7. Crosses of California and other firs (accompanied by successful control crosses, except as noted).

Parents3 Pollen age No. of No. of No. germinable Female Male (yr) crosses cones seeds per cone concolor (2) amabilis (1) 3 2 9 0 concolor {2) firma (1) 0 2 5 0 concolor (\) forrestii (1) 0 14 0 concolor (3) lasiocarpa (1) 3 3b 15 0 concolor (1) pinsapo (1) 0 1 8 0 concolor (1) religiosa (1) 2 1 6 9.5 concolor (3) veitchii (1) 2 3 7 0 concolor (1) ve/«n (1) 4 1 6 0 forrestii (2) concolor (2) 3 2C 8 0 forrestii {2) grandis (2) 3 2C 13 0 magnified (2) amabilis (1) 2 2 9 0 magnifica (2) lasiocarpa (2) 2 2 32 0 magnifica (1) religiosa (1) 2 1 5 0 magnifica (1) vejari (\) 4 1 10 0

a Number of trees in parentheses. b Controls lacking in one cross. c Controls lacking.

(1971). Within the few groups that have been studied, crossability is high, but crosses between groups are relatively difficult and sometimes impos sible. Three North American groups (sections of Liu 1971) are now known to consist mostly (Grandes) or entirely (Balsameae, Nobiles) of taxa that can easily be hybridized. In all three sections, species-pairs joined by zones of intergradation have proved to be moderately to highly crossable:

1. Balsameae: The eastern elements of this group were fully crossable except for combinations of balsam fir females and males of its bracted variety (A. balsamea var. phanerolepis), which averaged 47% crossability (Hawley and DeHayes 1985a). None of the eastern taxa were fully compatible with subalpine fir, but overall they averaged about 50% crossability with this western species. 2. Nobiles: Crossability of noble and red firs was 100% with red fir as female parent but only 29% with noble fir as female parent (Silen et al. 1965). Four noble firs but only one red fir were used as pollen parents in the control crosses on which crossability estimates are based, so 29% may be a closer estimate of the crossa bility of these Pacific Slope taxa. 3. Grandes: Crossability was high (47% to 86%) among elements of the white fir- grand fir complex, but crosses between the California and Rocky Mountain races of white fir averaged only 60% as much germinable seed per cone as crosses within the two geographic races. Interracial breeding barriers are un common in conifers, and this estimate is based on limited data. If a partial bar rier exists, however, white fir has a parallel in Pinus ponderosa Laws.: Cali fornia and Rocky Mountain races of this pine have a crossability of about 35% (Critchfield 1984a).

Another species in sect. Grandes, Pacific silver fir, has been reported to cross with white fir in Germany (Table 2), but our intra- and intersectional crosses on white and red fir females produced no germinable seeds (Table 7). Pacific silver fir grows with grand, subalpine, and noble firs in the Pacific Northwest, but no natural hybrids have been verified. This limited evidence of genetic isolation suggests that Pacific silver fir may not be closely related

148/ Forest Science to either the white fir-grand fir complex or the other northwestern firs, and may not belong in sect. Grandes. In contrast to generally high crossability of taxa in the same section, most crosses between taxa in different sections were unsuccessful. A total of 29 crosses in both directions between red fir (sect. Nobiles) and the grand fir- white fir group (sect. Grandes) produced a few probable hybrid seedlings that failed to survive their first growing season. The strong reproductive barrier between red and white firs is reinforced in nature by different polli nation times, and no natural hybrids have ever been reported. Other evi dence of the separateness of sect. Nobiles is the failure of controlled crosses between red fir and five other North American firs (Tables 5 and 7). Attempts to cross bristlecone fir were restricted to two other California species, red and white firs. The only products of this effort were two short lived white x bristlecone fir hybrids. Bristlecone fir is so different morpho logically from other firs that it is usually segregated in a separate subgenus, and these preliminary crossing results suggest that its isolation from other firs may be nearly complete. The only intersectional cross that yielded numerous healthy hybrids was the combination of Rocky Mountain white fir and oyamel (St. Clair and Critchfield, in prep.), one of the Mexican firs that constitute sect. Oyamel (Table 1). Other vigorous hybrids between taxa in different taxonomic groups were produced by Hawley and DeHayes (1985a) from crosses be tween balsam and Fraser fir females and white fir pollen parents. Their re ciprocal crosses failed, however, as did our crosses of white fir females and subalpine fir pollen parents (Table 7). All of these crosses are intersectional according to the classification of Liu, but white fir x oyamel is a cross be tween members of two groups in the same section of Gaussen's classifica tion, and Franco placed white fir and balsam, Fraser, and subalpine firs in different series of the same section (Table 1). Although crosses between North American sections are sometimes suc cessful, firs of the Eastern and Western hemispheres may be isolated from each other by genetic barriers that are almost or completely insurmount able. The five interhemisphere combinations we attempted (one in both di rections) were all failures (Table 7). Six other failed combinations involving California firs have been described by others (Table 3). An interhemisphere combination not involving a California fir, Kormutak's (1981) repeated at tempts to cross A. balsamea x A. alba, failed due to post-fertilization col lapse. On the other hand, at least 15 combinations involving four Western Hemisphere species and eleven Eastern Hemisphere species have been claimed as successes. They include: (a) nine Unverified hybrids (Table 2), among them three combinations described as failures by others (Table 3); (b) two unverified hybrids not involving California species: A. procera x A. mariesii Masters, and A. procera x A. sachalinensis (Mergen et al. 1964); and (c) four combinations of Uncertain status (Table 2), for which available descriptive data are sketchy and inconclusive. Any of these groups could include authentic hybrids, but in the absence of a single adequately de scribed interhemisphere hybrid it can be tentatively concluded that most or all combinations of Eastern and Western hemisphere firs are difficult or im possible to hybridize. Gaussen's classification differs notably from those of Liu and Franco in placing together Eastern and Western hemisphere firs in the same section of the genus (Table 1). Of the five groups in Gaussen's Section I, all with me dial resin canals in their leaves, the first is equivalent to Liu's North Amer ican sect. Balsameae (the only American firs with medial canals) and the

March 1988/ 149 other four include only Asian species. Two of the putative hybrids listed by Mergen et al. (1964), A. sachalinensis x A. lasiocarpa and A. koreana x A. lasiocarpa, are crosses between Asian and North American groups in Gaussen's section. If these putative interhemisphere hybrids still survive, their study would provide a useful test of Gaussen's unorthodox classifica tion. In its crossing barriers between Eastern and Western hemisphere species, Abies resembles the more thoroughly explored hard pine group (sect. Pinus of the genus Pinus). Despite crossing attempts on a much larger scale than in Abies, no verified hybrids have been produced between hard pines native to Eastern and Western hemispheres (Critchfield 1986). Among the white pines of sect. Strobus, by contrast, interhemisphere crosses are common, with crossabilities ranging up to 40%. The genus Picea, although it re sembles Abies in the nearly complete absence of the self-defining groups that characterize Pinus, also contrasts with Abies in the high crossability of some interhemisphere combinations—44% for crosses of the North Amer ican black spruce {P. marina [Mill.] B.S.P.) and Serbian spruce (P. omorika [Pancik] Purkyne) of southeastern Europe (Fowler 1980). Although our knowledge of the crossing behavior and taxonomy of the firs is still in a rudimentary state compared with what is known about the spruces or pines, it is nevertheless certain that the view of the firs as a group virtually lacking in crossing barriers is no longer tenable.

LITERATURE CITED

Critchfield, W. B. 1977. Sargent's fir hybrid: Abies amabilis x lasiocarpa. J. Arnold Arbor. 58:52-59. Critchfield, W. B. 1984a. Crossability and relationships of Washoe pine. Madrono 31:144- 170. Critchfield, W. B. 1984b. Impact of the Pleistocene of the genetic structure of North Amer ican conifers. P. 70-118 in Proc. 8th North Am. For. Biol. Workshop, R. M. Lanner (ed.). Logan, UT. Critchfield, W. B. 1986. Hybridization and classification of the white pines (Pinus section Strobus). Taxonomy 35:647-656. Cumming, W. C, and F. I. Righter. 1948. Methods used to control pollination of pines in the Sierra Nevada of California. USDA Circ. 792.18 p. Fowler, D. P. 1980. Hybridization of black spruce and Serbian spruce. Can. For. Serv., Mari- times For. Res. Cent. Inf. Rep. M-X-112. 30 p. Franco, J. d. A. 1950. Abetos. An. Inst. Super. Agron. 17. Lisbon, Portugal. 260 p. Franklin, J. F. 1961. A guide to seedling identification for 25 conifers of the Pacific North west. USDA For. Serv. Pac. Northwest For. & Range Exp. Stn., Portland, OR. 65 p. Gaussen, H. 1964. Les gymnospermes actuelles et fossiles, Fasc. VII. Genres Pinus (suite), Cedrus et Abies. Toulouse, Fr. P. 273-480. Griffin, J. R., and W. B. Critchfield. 1976. The distribution of forest trees in California. USDA For. Serv. Res. Pap. PSW-82 (rev.). 118 p. Hamrick, J. L., and W. J. Libby. 1972. Variation and selection in western U.S. montane species I. White fir. Silvae Genet. 21:29-36. Hawley, G. J., and D. H. DeHayes. 1985a. Hybridization among several North American firs I. Crossability. Can. J. For. Res. 15:42-49. Hawley, G. J., and D. H. DeHayes. 1985b. Hybridization among several North American firs II. Hybrid verification. Can. J. For. Res. 15:50-55. Kantor, J., and E. Chira. 1971. On the possibility of crossing certain species of the genus Abies. Acta Univ. Agric. (Brno), Ser. C (Fac. silv.) 40(1): 15-27.

150/ Forest Science Klaehn, F. U., and J. A. Winieski. 1962. Interspecific hybridization in the genus Abies. Silvae Genet. 11:130-142. Kormutak, A. 1981. Cytological study on crossability of some Abies species. Biologia (Bra tislava) 36:245-251. Kormutak, A., and J. Dubovsky. 1984. Embryo inviability in Abies concolor var. lowiana x Abies alba crossing. Biol6gia (Bratislava) 39:3-14. Larsen, C. S. 1934. Forest tree breeding. P. 93-113 in Roy. Vet. & Agric. Coll. Copenhagen Yearb. Ledig, F. T. 1987. Genetic structure and the conservation of California's endemic and near-en demic conifers. P. 587-594 in Conservation and management of rare and endangered plants T. S. Elias (ed.). Proc. Calif. Native Plant Soc. Symp., Sacramento, CA. Liu T. S. 1971. A monograph of the genus Abies. Dep. For. Nat. Taiwan Univ., Taipei, Taiwan. 608 p. Mergen, E, J. Burley, and B. A. Simpson. 1964. Artificial hybridization in Abies. Der Zuchter 34:242-251. Rohmeder, E. 1961. Praktische Anwendungsmoglichkeiten forstgenetischer Forschungs- ergebnisse. Schweiz. Z. Forstwes. 112:43-71. Silen, R. R., W. B. Critchfield, and J. F. Franklin. 1965. Early verification of a hybrid between noble and California red firs. For. Sci. 11:460-462. St. Clair, J. B., and W. B. Critchfield. 198 . Hybridization between a Rocky Mountain and Mexican fir: Abies concolor x religiosa. In prep. Wright, J. W. 1953. Summary of tree-breeding experiments by the Northeastern Forest Ex periment Station 1947-1950. USDA For. Serv. Northeast. For. Exp. Stn. Pap. 56. 47 p. Zavarin, E., W. B. Critchfield, and K. Snajberk. 1978. Geographic differentiation of monoterpenes from Abies procera and Abies magnifica. Biochem. Syst. & Ecol. 6:267-278. Zavarin, E., K. Snajberk, and W. B. Critchfield. 1977. Terpenoid chemosystematic studies of Abies grandis. Biochem. Syst. & Ecol. 5:81-93.

March 1988/ 151