Summary Surrogate pollen induction (SPI) was evaluated on loblolly pine (Pinus taeda L.) donor scions from 5-year-old progeny that were grafted by topworking into the lower crowns of 16-year-old loblolly pine receptor clones in a seed orchard. On each of 25 study trees, one of three pollen induction treatments (wire girdle, saw girdle or control) was applied to 10 receptor branches below the graft location. Graft survival was 76%. Of the surviving grafts, 57% produced pollen strobili in March 1993, 13 months after grafting. The pollen induction treatments did not decrease graft survival or increase pollen production. Graft survival did not vary significantly among the donor scion genotypes, but the percentage of grafts with pollen was significantly related to the donor scion genotype. The mean number of pollen clusters induced per ramet also differed significantly among the donor scions. There was a tendency for pollen phenology of the grafted scions to be modified by the receptor clone. We conclude that surrogate pollen induction, coupled with accelerated female flower stimulation, can reduce the breeding schedule in loblolly pine to 3 years.
Keywords: forest genetics, grafting, pollen induction, topwork-ing, tree breeding.
Introduction
The long generation interval of loblolly pine remains an obsta-cle in breeding programs for rapid genetic improvement. When scions are collected from reproductively mature trees and grafted on 1- to 2-year-old rootstocks, male and female strobi-lus production normally begins in 3 to 4 years. Pollen produc-tion invariably lags behind female strobilus producproduc-tion, delaying completion of the entire selection and breeding cycle. This time lag may be 5 to 8 years in seedlings, 3 to 4 years in seed orchard grafts and 1 to 2 years in accelerated breeding programs (Schmidtling and Greenwood 1993, Greenwood 1995).
Burris et al. (1991) developed an abbreviated breeding schedule. They applied flower stimulation treatments to lob-lolly pine (Pinus taeda L.) in the same year that the scions were grafted on juvenile stock and showed that female flowers could be produced in a greenhouse 14 months after grafting scions from 3- and 8-year-old plants, but not after grafting scions from 1-year-old seedlings. By 26 months after grafting, scions
from 1-, 3- and 8-year-old plants produced female flowers. No pollen strobili were produced on any grafts 14 months after grafting; however, after 26 months, scions from 1-, 3- and 8-year-old plants produced pollen, and there was no detectable effect of scion age on the number of pollen clusters per graft.
The abbreviated breeding schedule developed by Burris et al. (1991) reduced the time required for female strobilus pro-duction, but because no pollen was produced until 26 months, the breeding schedule required 4 years to complete (Fig-ure 1B). Nevertheless, their approach was an improvement over the 5-year breeding schedule (Figure 1A) proposed by Greenwood et al. (1986). To reduce the breeding schedule further, pollen would need to be produced within 1 year from selection (Figure 1C). The objective of our study was to evalu-ate a grafting procedure, surrogevalu-ate pollen induction (SPI), in which young scions were grafted onto older trees to induce pollen in the same year as grafting.
Greenwood and Gladstone (1978) used a grafting proce-dure, termed topworking, that involved grafting 1-year-old seedlings into the upper crowns of mature seed orchard trees. Two and 3 years after grafting, 10% of the surviving grafts produced pollen, and 3 and 4 years after grafting, 59% of the topworked grafts produced pollen.
Our surrogate pollen induction method differed from pre-vious procedures in that the scions were 5 years old, the current age of selection in loblolly pine, and we grafted the donor scions into the lower crowns of receptor seed orchard clones that were producing abundant pollen. The receptor clones were large interstocks that had been grafted for commercial seed production.
Materials and methods
Genetic material
Five ramets each from five grafted loblolly pine clones in the Weyerhaeuser Company Seed Orchard at Lyons, Georgia (32° lat., 82° long.), were selected as receptor clones in late winter of 1992. The receptor clones were 8 years old at the time of the surrogate pollen induction treatments. The seed orchard is intensively managed, and practices include mowing, herb-aceous weed control, fertilization and insect protection. All receptor clones produced pollen in 1992.
Surrogate pollen induction shortens the breeding cycle in loblolly pine
D. L. BRAMLETT,
1C. G. WILLIAMS
2and L. C. BURRIS
31 USDA Forest Service, Route 1, Box 182A, Dry Branch, GA 31020, USA
2 Department of Genetics, North Carolina State University, Raleigh, NC 27695, USA 3 Weyerhaeuser Co., Hot Springs, AR 71902, USA
Received May 31, 1994
Donor scions were collected in February 1992 from five individuals in each of five unrelated full-sib families that had superior growth in an Arkansas second-generation genetic test (Williams and Lambeth 1995). The scions were 5 years old. Grafting was completed in February 1992.
Ten lower branches were selected on each ramet, and one of three pollen induction treatments was randomly applied in February 1992 several days before grafting. The branch treat-ments included: (1) wire girdling, (2) saw girdling, and (3) an untreated control. Three branches were wire girdled, three branches were saw girdled, and four branches per ramet were left as untreated controls.
A secondary branch on the numbered primary branch was selected, which was approximately the same diameter as the scion. The scions were side grafted to the secondary branch and protected with rubber strips and melted wax (80--90 °C) until shoot growth of the scion began. Surviving grafts were released by pruning the receptor shoot above the graft in one or two stages in May and June 1992.
Experimental design
The experimental design was a split-plot with receptor clones as blocks, individual ramets as whole plots, and randomized branch treatments as subplots. Because branch treatments had no effect on graft survival or pollen production, the 10
branches were pooled. Subsequent analyses of variance with a randomized complete block design model were conducted with the GLM procedure of SAS (SAS Institute Inc., Cary, NC). Seed orchard receptor clones were considered a random effect and selected donor scions were a fixed effect. Means for donor scions were separated with Tukey’s studentized range test. In the split-plot ANOVA, branch treatment effects (2 df) were tested with the pooled interactions receptor clone × branch treatment (8 df) and receptor clone × selected scion × branch treatment (32 df) as the error term (Table 1). Once the branch treatment was dropped from the analysis, the statistical model became a randomized complete block design (Table 2). The donor scion effect (4 df) was tested against the residual error mean square (16 df).
Response variables
Response variables were: (1) graft survival = the number of live grafts divided by the total number of grafts completed; (2) live grafts with pollen = the number of live grafts with pollen divided by the number of live grafts; and (3) pollen cluster ratio = the number of clusters of pollen on live grafts divided by the total number of grafts completed. There were 50 grafts for each donor scion source and 10 grafts per individual receptor clone. A total of 250 grafts were completed on the 25 receptor ramets. Pollen phenology
The reproductive phenologies of all donor scions with pollen clusters were scored on March 4, 18 and 25, 1993, according to the pollen development classification system (PDCS) de-scribed by Bramlett and Bridgwater (1989). Based on this system, a six-stage scale describes loblolly pine pollen phe-nological development: Stage 3 is the period of rapid strobilus elongation before pollen shedding, Stage 4 is the start of pollen release, Stage 5 is the period of maximum pollen release, and
Figure 1. Breeding schedules in loblolly pine for conventional indoor breeding (top) (Greenwood 1986), abbreviated breeding (middle) (Burris et al. 1991), and surrogate pollen induction breeding (bottom).
Table 1. ANOVA for split-plot experimental design.
Source df
Receptor clone (RC) 4
Donor scions (DS) 4
RC × DS 16
Branch treatment (BT) 2
DS × BT 8
Error term
RC × BT 8
RC × DS × BT 32
Total 74
Table 2. ANOVA for randomized complete block experimental design.
Source df
Receptor clone (RC) 4
Donor scions (DS) 4
RC × DS 16
Stage 6 marks the end of pollen release. Stages 3 and 5 are further subdivided. The PDCS was also used to quantify the duration of pollen release and the phenological development for a given date. On each measurement date, adjacent pollen-bearing branches in each receptor clone were scored for PDCS. Independence between the PDCS scores from donor scions and receptor clones was tested with Spearmen’s rank correla-tion test for nonparametric data.
Results
Of the 250 grafts topworked into the seed orchard clones, 76% survived and produced shoot growth 1 year after grafting. Of the surviving grafts, 57% produced pollen. The number of pollen clusters produced per donor scion source was a function of both graft survival and the number of potential sites (vege-tative shoots). Individual grafts produced from zero to five pollen cone clusters with an average of 1.7 clusters per live graft. Donor scions had varying numbers of pollen clusters per ramet (10 grafted scions). Scion DR-3 produced an average of 12.8 clusters per ramet, which was significantly more than the 1.2 pollen clusters per ramet for DR-0 (Table 3). On average, the SPI technique resulted in the production of four to five clusters per 10 donor scion grafts after 13 months. A
subsam-ple of 20% of the pollen clusters averaged 5.8 strobili per cluster. Although the average number of vegetative shoots on the surviving grafts varied from 9.6 to 17.0 per ramet (10 branches), the differences were not statistically significant (Table 3).
The branch treatments of wire girdling and saw girdling had no effect on the number of surviving grafts, the number of vegetative shoots per graft, or the ratio of pollen clusters to attempted grafts (Table 4). The lack of branch treatment effect on pollen initiation indicates that the branches selected for grafting were already in an inductive state for male strobilus production and production was not enhanced or hastened by the girdling treatments. All of the remaining analyses were based on the pooled data of individual branches per ramet (10 branches).
Graft survival was not a function of the donor scion (Ta-ble 5). Although scion survival varied from 40 to 100% by scion source on individual ramets, there was no significant difference in graft survival by scion source across all seed orchard clones. One year after grafting, the number of surviv-ing grafts with pollen clusters was significantly affected by source of the scion material. Only 15% of the surviving scions from DR-0 produced pollen compared to 96% of the surviving scions from DR-3 (Table 5).
Table 3. Number of vegetative shoots and the number of pollen catkin clusters per ramet on surviving grafts of 5-year-old loblolly pine scions grafted on 8-year-old seed orchard clones.
Donor scion Seed orchard clone
2-4 8-5 8-6 11-3 11-5 Average1
Vegetative shoots per 10 grafts
DR-0 15 8 10 10 5 9.6 a
DR-1 18 18 14 14 8 14.4 a
DR-3 21 27 13 11 11 16.6 a
DR-4 28 13 20 19 5 17.0 a
DR-7 21 16 17 12 5 14.2 a
Total 103 82 74 66 34
Pollen clusters per 10 grafts
DR-0 1 0 3 2 0 1.2 b
DR-1 12 6 11 5 0 6.8 ab
DR-3 15 20 12 10 7 12.8 a
DR-4 8 7 18 5 0 7.6 ab
DR-7 19 11 14 10 3 11.4 a
Total 55 44 58 32 10
1 Means with the same letter are not significantly different (P = 0.05) using Tukey’s studentized range test.
Table 4. Graft survival, number of vegetative shoots per graft, and the ratio of pollen clusters per completed graft on 5-year-old loblolly pine scions grafted on 8-year-old seed orchard clones receiving three branch treatments.
Branch Survival1,2 Vegetative shoots/grafts1,2 Pollen clusters/graft1,2
treatment (%) (number) (ratio)
Wire girdle 77 1.41 0.79
Saw girdle 73 1.53 0.87
Control 76 1.36 0.75
1 Based on total grafts completed.
The results of the phenological observations were not clear. The PDCS scores of donor clone DR-3 appeared to remain constant regardless of the phenology of the receptor clone (Table 6), whereas the PDCS scores of donor clones DR-1 and DR-7 appeared to be influenced by the receptor clone. The Spearman rank correlation test was performed on 19 pairs of PDCS scores, and the null hypothesis of independence be-tween donor and receptor phenology was rejected at the 0.07 level of probability.
Discussion
The approaches for implementing SPI in genetic improvement programs for loblolly pine include: (1) grafting a relatively large number of scions (50 or more) and collecting all available pollen after 1 year to use in an accelerated breeding program; and (2) grafting relatively few scions per clone with the expec-tation that, after some delay in pollen production on some of the clones, most of the selections will have produced adequate amounts of pollen after 2 or 3 years. Our results indicate that 50 grafts from each selection would produce an adequate amount of pollen for a breeding program after 1 year for four out of the five donor scion sources (Bramlett and O’Gwynn 1981) (donor clone DR-0 produced only a marginal quantity of pollen).
There was a positive effect of SPI on pollen production in loblolly pine. It appears that buds of the grafted scions re-sponded to a pollen induction stimulus that is localized in the lower crowns of the trees. The lack of a branch girdling effect on pollen induction indicates that the pollen induction stimulus was not translocated in the phloem from the roots to the buds or vegetative shoots; however, it is possible that a pollen induction stimulus was translocated in the xylem. Pollen in-duction occurred even though the buds on grafted scions were
formed on vegetative shoots that had small needles, a juvenile appearance and delayed vegetative development. It appears that pollen induction is a localized effect in the lower crown and occurs most frequently on secondary and tertiary branches with restricted shoot elongation. The 5-year-old scions that were grafted into the lower crown were reproductively compe-tent and initiated pollen strobili on 57% of the surviving grafts. Reduced shoot elongation could promote pollen induction by allowing the resting bud adequate time to differentiate primor-dia. In contrast, grafts on 1-year-old seedlings exhibited vigor-ous growth and multiple cycles of shoot elongation, and no pollen was produced on these grafts after 13 or 25 months; however, after 37 months, a small amount of pollen was pro-duced on a few grafts (authors’ unpublished observations).
Surrogate pollen induction offers the potential to reduce the breeding schedule of loblolly pine to 3 years. To accomplish a 3-year breeding schedule, it will be necessary to combine SPI with the techniques used by Burris et al. (1991) to stimulate female flower production 1 year after selection. Thus, if scions of the pollen parents were grafted in the lower crown of seed orchards that are already producing substantial amounts of pollen, pollen would be available 13--14 months after grafting and mature seed would be produced 18 months after pollina-tion. Not all scions and clones will have enough pollen and flowers after 1 year, but female flower and pollen production should increase in succeeding years.
Although we only evaluated a single scion aged 5 years, it seems likely that the positive effect of the crown location could induce pollen development in scions younger than 5 years old. Thus it may be possible to complete the entire breeding cycle on very young material. Furthermore, if scions from selected genotypes can be induced to form female strobili in the upper crown of receptor clones, the entire breeding process could be completed on surrogate interstocks.
Table 5. Percentage of surviving grafts 1 year after grafting and the percentage of live grafts with pollen clusters per ramet on 5-year-old loblolly pine scions grafted on 8-year-old seed orchard clones.
Donor scion Seed orchard clone1
2-4 8-5 8-6 11-3 11-5 Average2
1 Ten grafts were completed for each scion source on each seed orchard clone.
Acknowledgment
This research was supported by a cooperative agreement between the USDA Forest Service and Weyerhaeuser Company. The authors ap-preciate the support of Franklin Brantley and the staff at Weyer-haeuser’s Lyons Seed Orchard, Lyons, Georgia.
References
Bramlett, D.L. and F.D. Bridgwater. 1989. Pollen development classi-fication system for loblolly pine. Proc. 20th South. For. Tree Im-prov. Conf., Charleston, SC, pp 116--121.
Bramlett, D.L. and C.H. O’Gwynn. 1981. Controlled pollinations. In
Pollen Management Handbook. Ed. E.C. Franklin. USDA For. Serv. Agric. Handbook 587, pp 44--57.
Burris, L.C., C.G. Williams and S.D. Douglas. 1991. Flowering re-sponse of juvenile selections in loblolly pine. Proc. 21st South. For. Tree Improv. Conf., Knoxville, TN, pp 110--119.
Greenwood, M.S. 1995. Recent advances in the control of reproduc-tive development of conifers. In Proc. IUFRO Conf. Tropical Tree Breeding, Cali, Columbia. In press.
Greenwood, M.S., C.C. Lambeth and J.L. Hunt. 1986. Accelerated breeding and potential impact upon breeding programs. LA Agric. Exp. Stn., South. Coop. Serv. Bull. 309, pp 39--43.
Greenwood, M.S. and W.T. Gladstone. 1978. Topworking loblolly pine for precocious flowering. Weyerhaeuser Co. Tech. Rep. 042-3004/78/80, Hot Springs, AR, 8 p.
Schmidtling, R.C. and M.S. Greenwood. 1993. Increasing pollen production. In Advances in Pollen Management. Ed. D.L. Bramlett, G.R. Askew, T.D. Blush, F.E. Bridgwater and J.B. Jett. USDA For. Serv. Agric. Handbook 698, pp 33--39.
Williams, C.G. and C.C. Lambeth. 1995. Genetic improvement using an elite breeding population. In Proc. IUFRO Conf. Tropical Tree Breeding, Cali, Columbia. In press.
Table 6. Mean March 25, 1993 Pollen Development Classification System (PDCS) scores observed on seed orchard clones and topworked scions from 5-year-old loblolly pine.
Donor scion Seed orchard clone Donor scion
Number of observations PDCS score1 Number of observations PDCS score1
Receptor clone 2-4
DR-0 10 3.60 0
--DR-1 10 3.60 12 3.65
DR-3 14 3.56 15 3.64
DR-4 9 3.60 8 3.68
DR-7 11 3.38 9 3.64
Receptor clone 8-5
DR-0 9 4.56 0
--DR-1 9 4.96 6 3.77
DR-3 10 4.44 20 3.65
DR-4 13 5.45 7 5.05
DR-7 12 -- 11 --
Receptor clone 8-6
DR-0 9 5.22 3 3.60
DR-1 14 5.17 11 4.09
DR-3 12 4.97 12 3.73
DR-4 15 5.08 18 5.19
DR-7 11 3.81 13 3.60
Receptor clone 11-5
DR-0 16 3.74 0
--DR-1 12 3.60 0
--DR-3 11 4.61 7 3.60
DR-4 12 3.72 0
--DR-7 14 4.21 3 3.60
Receptor clone 11-3
DR-0 15 5.78 2 3.60
DR-1 10 5.80 5 4.90
DR-3 8 5.15 10 3.64
DR-4 10 5.68 5 3.92
DR-7 12 5.79 10 4.46
1 At PDCS Stage 3.0, strobili begin to increase in length and exude a milky fluid when pressed. At Stage 3.3, strobili continue elongation and
