Summary In 1989, identical crosses (2--3 females within males) were performed with Picea abies (L.) Karst. in a green-house seed orchard at Biri nursery and in an outdoor seed orchard at Huse, 32 km north of Biri. Pollination began 17 days earlier in the greenhouse than outdoors at Huse. The potted grafts in the greenhouse were moved outdoors when the seed cones were no longer receptive. Twelve full-sib family pairs (Biri and Huse) from these crosses were grown in a phytotron and tested for height and autumn frost hardiness during their first growing season. No significant difference was found be-tween the indoor (Biri) and outdoor (Huse) progenies for height growth. However, the progenies from the greenhouse seed orchard were significantly more susceptible to frost than their full-sibs from the outdoor seed orchard. There was no signifi-cant interaction between males and the flowering environment, but a significant female × flowering environment interaction was present as a result of greater differences in frost hardiness between progenies from females in the greenhouse seed or-chard than in the outdoor seed oror-chard. Although seeds from the outdoor seed orchard generally had a greater biomass than seeds from the greenhouse seed orchard, the difference in seed weight did not explain the difference in frost hardiness. We hypothesize that temperature and photoperiod during pollina-tion and fertilizapollina-tion affect the frost hardiness of the progenies. Keywords: climate, flowering, height growth, photoperiod, pollination, seed biomass, temperature.
Introduction
There is growing interest in the use of indoor containerized seed orchards for acceleration of both breeding and production of genetically improved tree seeds (Ross et al. 1986, Philipson 1990). Containerized orchard designs permit better control of the flowering process and give greater flexibility in choice of genotypes to be interbred than outdoor seed orchards. In 1983, a greenhouse seed orchard was established with Picea abies (L.) Karst. at Biri nursery in Norway. Flowering and seed production in response to heat and gibberellin A4/7 treatments have generally been good in the greenhouse (Johnsen et al. 1994a, 1994b).
Recent results with Norway spruce and Scots pine have shown that the parental environment, or more specifically, the climate during sexual reproduction, may influence the adap-tive properties of the progeny (Johnsen 1988, 1989a, 1989b, Johnsen et al. 1989, Dormling and Johnsen 1992, Johnsen and Østreng 1994, Skrøppa 1994, Skrøppa et al. 1994). Such ef-fects have been observed when northern parents have been grafted in a southern seed orchard or parents from high alti-tudes have been transferred to orchards located at low eleva-tions. Thus, the seedlings from these parents have an extended growth period and a delayed development of frost hardiness in the autumn relative to what would be expected. Although the altered phenotypic performance is long-lasting in Norway spruce (Johnsen 1989a, 1989b, Skrøppa 1994), it may be only temporary for autumn frost hardiness in Scots pine (Dormling and Johnsen 1992). However, field trials have shown that both mortality and growth of Scots pine between 4 and 5 years from seed are affected by the parent plant environment, and that the southern localities produced progenies with the highest mor-tality (Lindgren and Wei 1994). In the greenhouse seed orchard at Biri, crosses are normally performed inside the greenhouse, but the grafts are moved outside shortly after the female flow-ers are no longer receptive. Thus, the female grafts are grown at an elevated temperature for approximately 3 weeks in spring. In this study, we investigated whether this short period inside the greenhouse could affect progeny performance.
Materials and methods Crosses performed
Identical crosses were performed on P. abies grafts in a green-house at Biri nursery (60°57′ N, 10°33′ E, 150 m a.s.l.) and in an outdoor seed orchard at Huse (61°13′ N, 10°20′ E, 180 m a.s.l.), located 32 km north of Biri. Five male and 12 female parents were crossed according to a mating design with 2--3 females within males (Figure 1). The parents originated from southeastern Norway (60--61° N, 160--330 m a.s.l.) and were unrelated. The grafts at Biri were grown in 50-l pots and were approximately 2.5--3 m tall when pollination took place. The grafts at Huse were 10--13 m tall when the crosses were
Sexual reproduction in a greenhouse and reduced autumn frost
hardiness of
Picea abies
progenies
ØYSTEIN JOHNSEN,
1TORE SKRØPPA,
1GUNNAR HAUG,
2INGER APELAND
1and GEIR ØSTRENG
11 Norwegian Forest Research Institute, Division of Silviculture, Høgskoleveien 12, 1432 Ås, Norway
2 The State Forest Seed Station, P.O. Box 118, 2301 Hamar, Norway
Received May 31, 1994
performed. The grafts at Biri nursery had been treated with gibberellin and heat in 1988 (Johnsen et al. 1994a).
The males were moved to the greenhouse during the last week of April 1989, and mature pollen cones were collected from each clone separately during the first week of May. The day/night temperature in the greenhouse was 20--25/15 °C. No artificial light was supplied. Pollen was extracted in the labo-ratory at 20--22 °C, dried and cleaned. Approximately 7--8 ml samples of pollen were placed in 15-ml glass vials and sealed. Pollen from the greenhouse was used in crosses at both Biri and Huse.
Paper bags were used to isolate female flower buds before opening. Buds were isolated in the last week of April at Biri and in the first week of May at Huse. At Biri, the potted female clones were moved inside the greenhouse during the last week of April, and crosses were performed from May 5 to 11. The day/night temperature inside the greenhouse was 20--25/15--20 °C. The female flowers developed quickly, and the paper bags were opened on May 16. The female grafts were moved outside 3 days later. At Huse seed orchard, the female flowers were not receptive until the third week of May as a result of cold weather during the first 2 weeks of May. Thus, pollination took place 17 days later at Huse than at Biri (from May 22 to 25). The mean temperature during the week of pollination at Huse varied from 7 to 16 °C, and the paper bags were opened the first week of July.
At both Biri and Huse, mature cones were collected in September 1989, and seeds were extracted, dewinged, cleaned and separated from empty seeds, and the weight of approxi-mately 200 seeds was recorded per seed lot.
Cultivation of plants
Twelve pairs (each pair comprised one seed lot from Biri and one seed lot from Huse) of full-sib families were sown the first week of November 1991 in two rooms at the Phytotron at the University of Oslo. Seedlings were grown in multipot contain-ers in continuous light (270 µmol m−2 s−1) at 22 °C until they were 8--10 cm tall (second week of January 1992) and then gradually induced to harden as described previously (Johnsen 1989c). Briefly, the hardening program started with 3-h nights for 1 week followed by 1 week with 6-h nights. The night
length was then increased by 1 h per week. When the length of the night reached 9 h, the seedlings were exposed to a night temperature of 10 °C for 1 week. Thereafter, the night tempera-ture was set at 5 °C.
Freezing tests
Plants were harvested for the first freezing test when the night length reached 11 h in the phytotron (March 5). The freezing test temperatures were −6, −7.5, −9, −10.5 and −12 °C. The freezing test was started by subjecting seedlings to a tempera-ture of 5 °C for 1 h; the seedlings were then frozen at a rate of 2 °C h−1 to the test temperature, held at the test temperature for 4 h, then thawed at a rate of 2 °C h−1 to 5 °C, and finally held at 5 °C for 3 h. Plants were misted for 15 s at 0 °C during freezing. The second freezing test was performed on March 11 (12-h night in the Phytotron) under the same conditions as described for the first test. The roots were insulated during freezing. After freezing, the plants were grown in a heated greenhouse in continuous light at a day/night constant tem-perature of 22 °C for 3 weeks. Frost damage on needles (browning and discoloration) was assessed based on a scale from 0 to 11 where: 0 = no visible damage; 1--10 = percent increments of brown or discolored needle length; and 11 = all needles completely brown (Johnsen 1989c). Plant height was measured when the seedlings were scored for frost damage. Experimental design and statistical analyses
During growth and cold acclimation, the plants were grown in two Phytotron rooms with two replicates (complete blocks) per room. The plants were grown in contiguous split-plots (n = 40) with each pair of full-sib families (Biri and Huse) in one container per replicate. Thus, a replicate was comprised of 12 containers (one for each pair of full-sibs).
Plants were harvested, sorted and placed in new containers at the start of the freezing tests. In each freezing chamber, each pair of families was replicated four times, and the replicates were arranged to correspond with the four replications during plant cultivation, i.e., two seedlings from each member of the family pairs that had grown in adjacent plots within Phytotron rooms, were also placed in adjacent plots within chambers during freezing. The data from the test temperatures that gave a mean frost damage score of 3.5--8.5 were used in the analyses of variance. These test temperatures were −6 and −7.5 °C from Test 1 on March 5, and −7.5, −9 and −10.5 °C from Test 2 on March 11.
The analyses of variance of height and frost damage were performed with the GLM procedure (SAS Institute Inc., Cary, NC) according to Model 1:
Yijklmno=µ+Li+ Fj+M(j)k+LFij+LMi(j)k+Tl+ C(l)m+PRn+R(n)o ,+eijklmno
where Yijklmno is the mean of two plants for each combination of the experimental factors for untransformed heights and the transformation arcsin(Xijklmno /11)1/2 for frost damage (X = mean of two plants; transformation homogenizes variances among groups and normalizes the data (Norell et al. 1986, Figure 1. Crosses performed at an indoor (Biri) and outdoor (Huse)
Johnsen 1989c)), µ is the total mean, Li is the fixed effect of seed location i where i = 1 or 2 (Biri or Huse), Fj is the random effect of father j where j = 1, 2, 3, 4 or 5, M(j)k is the random effect of mother k within father j where k = 1, 2 or 3, LFij is the interaction between father and seed location, LMi(j)k is the interaction between seed location and mother within father, Tl is the fixed effect of test temperature l where l = 1, 2, 3, 4 or 5, C(l)m is the random effect of freezing chamber m within tem-perature l where m = 1 or 2, PRn is the random effect of Phytotron room n where n = 1 or 2, R(n)o is the random effect of replicate o within room n where o = 1 or 2, and eijklmno is the experimental error.
The mean squares for the interaction between seed location and fathers (MSL×F) should normally be used as the error term when testing the effect of seed location. However, the variance component father × seed location (σL×F2) appeared negative for both height and frost damage when the analyses were run (MSL×F were smaller than residuals). Theoretically, a variance component cannot be negative, and σL×F2 was thus set to zero. The appropriate error term for seed location thus became the mean square for the interaction between seed location and mothers within fathers (MSL×M/F). For the same reason, the effect of fathers, which could be tested with the linear combi-nation MSM/F + MSL×F− MSL×M/F as the error term, was instead tested against MSM/F. Mothers were tested with MSL×M/F as the error term. The technical components (test temperature, cham-bers, Phytotron room and replicates) were specified to provide an indication of each source of variation.
Model 2 for seed weight had the same structure but was simpler than Model 1:
Yij=µ+Li+Fj+M(j)k+LFij+eij.
For seed weight, MSL×F was greater than the residuals and was used as a denominator when testing the effect of seed location.
Results
Except for mother Number 11, the seeds from Biri had less biomass than the seeds from Huse (Figure 2). Only seed location had a significant effect on seed biomass (Table 1).
The difference in plant height between the two localities was not statistically significant (Table 2, Figure 3). The difference between mothers and the interaction between mothers and seed location were both significant. The differences between mothers were larger at Biri for fathers Numbers 1 and 4, but not for mothers within fathers Numbers 2, 3 and 5 (Figure 3). These differences explain the significant interaction between mothers and seed location on plant height.
Figure 2. Weight of seeds produced in an indoor (Biri) and an outdoor (Huse) seed orchard.
Table 1. The ANOVA of seed weight (mg).
Source df MS F-Value P-Value
Seed location (L) 1 11.28 10.48 0.0317
Father (F) 4 0.95 0.57 0.6933
Mother/F (M/F) 7 1.65 2.52 0.1233
L×F 4 1.07 1.64 0.2666
Error 7 0.66
Table 2. The ANOVA of plant height (mm) and frost damage (transformed values).
Source df Height Frost damage
MS F-Value P-Value MS F-Value P-Value
Seed location (L) 1 3472.4 1.44 0.2692 2.87 16.88 0.0045
Father (F) 4 28309.4 2.67 0.1217 2.50 1.06 0.4423
Mother/F (M/F) 7 10607.9 4.41 0.0344 2.35 13.50 0.0014
L×F 4 686.9 1.05 0.3817 0.03 0.56 0.6946
M/F×L 7 2404.2 3.67 0.0007 0.17 2.93 0.0049
Temperature (T) 4 1617.2 2.29 0.1935 12.80 94.44 0.0001
Chamber/T 5 705.1 1.07 0.3727 0.14 2.28 0.0449
Room (R) 1 60776.4 14.45 0.0627 0.06 0.87 0.4496
Replicate/R 2 4204.6 6.41 0.0017 0.07 1.16 0.3150
Progenies from Biri were significantly less frost hardy than their full-sibs from Huse (Figure 4, Table 2). There was no significant interaction between males and seed location, but the interaction between females and seed location was signifi-cant (Table 2). The difference in frost hardiness between fe-males within fe-males was greater in progeny from the indoor seed orchard than from the outdoor seed orchard for fathers Numbers 1, 2, 4 and 5 (Figure 4), and explains the significant interaction between mothers and seed location.
Discussion
Progeny from the greenhouse seed orchard was less frost hardy than progeny from the outdoor seed orchard. Previously, we found that progenies of P. abies from a northern site in Finland were more hardy than their full-sibs from a southern site (Skrøppa et al. 1994). Together these two studies provide evidence for aftereffects of parent plant environment on Nor-way spruce progeny (Bjørnstad 1981, Johnsen 1988, 1989a, 1989b, Johnsen et al. 1989). In studies where pairs of half-sibs from contrasting environments were compared, it was found that aftereffects are long-lasting and affect phenology, seasonal growth rhythm and development of autumn frost hardiness (Bjørnstad 1981, Johnsen 1988, 1989a, 1989b, Johnsen et al. 1989). Alternative hypotheses, such as phenotypic selection of plus-trees (Johnsen and Østreng 1994) and indirect effects of seed weight cannot explain the altered phenotypic perform-ance (Bjørnstad 1981, Skrøppa 1988, Johnsen 1989a, Dorm-ling and Johnsen 1992).
Seed from the outdoor seed orchard had greater biomass than seed from the indoor seed orchard. Thus, in our study, seeds with the larger biomass produced the more frost hardy plants (Figures 2 and 4). In contrast, in other studies, the southern and warmer localities produced heavier seeds and less frost hardy plants (Bjørnstad 1981, Andersson 1985, Johnsen 1989a, Dormling and Johnsen 1992, Johnsen and Østreng 1994). Based on our findings, the hypothesis that heavy seeds delay bud set and reduce autumn frost hardiness has to be rejected (Skrøppa 1988). We conclude that the envi-ronmentally induced changes in seed weight cannot explain the environmentally induced change in autumn frost hardiness. Pollination occurred 3 weeks earlier in the greenhouse seed orchard than in the outdoor seed orchard. Because of the large difference in temperature between the greenhouse and the outdoors during pollination, seed cones passed through the different stages of receptiveness (Brøndbo 1971) more quickly in the indoor seed orchard than in the outdoor seed orchard. During the period in the greenhouse, it is likely that pollen tube growth took place, but we do not know if fertilization also occurred while the grafts were in the greenhouse. However, the grafts at Biri must have reached the various stages in the generative process (pollen tube growth, formation of archego-nia, egg development, fertilization, proembryo stages, degen-eration of embryos, and early embryo development) at least 3--4 weeks earlier than the grafts at Huse as a result of the 3-week period in the greenhouse. Thus, the prezygotic stages occurred in a 1.5 h shorter photoperiod at Biri than at Huse. We do not know whether this difference in photoperiod between the two sites was large enough to induce the observed after-effect on the progenies. However, similar results were found in a comparison of progenies from two contrasting seed years in Lyngdal seed orchard (Kohmann and Johnsen 1994). The spring was late in 1987, and pollination took place during the third week of May, whereas in 1989, pollination was 3 weeks earlier. The seeds from 1989 gave rise to seedlings that had a 1 h longer critical night length (delayed bud set) than seedlings from seeds produced in 1987. Possible mechanisms (genetic, epigenetic or physiological causes) to account for these results Figure 3. Mean height of seedlings from seed produced at an indoor
(Biri) and an outdoor (Huse) seed orchard for each pair of full-sib families.
have been discussed (Johnsen 1989a, Skrøppa and Johnsen 1994).
The finding that heavier seeds were produced at Huse than at Biri (Figure 2) indicates that the physiological status of the potted grafts at Biri differed from that of the large, vigorously growing grafts in the outdoor seed orchard at Huse. Both sets of grafts were grown outdoors during seed maturation with little difference in mean temperature between the two sites. However, the grafts at Biri had been induced to flower by treatment with gibberellin A4/7 the year before (Johnsen et al. 1994a), whereas the grafts at Huse flowered without hormonal treatment. Interestingly, Gutterman et al. (1975) showed that hormonal treatments given to parent plants of lettuce affected the performance of the progenies.
Plants from the present families are now being grown to test whether the differences in frost hardiness between plants from crosses at Biri and Huse will last for many growth seasons (Johnsen 1989a, 1989b, Johnsen et al. 1989, Lindgren and Wei 1994, Skrøppa 1994) or will only be temporary (Dormling and Johnsen 1992, Andersson 1994).
Acknowledgment
We thank the Nordic Forest Research Cooperation Committee for financial support, the staff at Biri nursery for kind help when crosses were performed, and the staff at the Phytotron, University of Oslo, for skillful handling of the plants.
References
Andersson, B. 1985. Freezing tests of northern Scots pine seedlings from seed orchards In Reforestation Material for Harsh Northern Sites. Ed. J.E. Nillson. Swed. Univ. Agric. Sci., Dept. For. Genetics Plant Physiol., Report No. 4, pp 75--83 (summary in English). Andersson, B. 1994. Aftereffects of maternal environment on autumn
frost hardiness in Pinus sylvestris seedlings in relation to cultivation techniques. Tree Physiol. 14:313--322.
Bjørnstad, Å. 1981. Photoperiodical after-effect of parent plant envi-ronment in Norway spruce (Picea abies (L.) Karst) seedlings. Medd. Nor. Inst. Skogforsk. 36:1--30.
Brøndbo, P. 1971. A pollination experiment in Norway spruce (Picea abies (L.) Karst.). Medd. Nor. Skogforsøksves. 111:65--77. Dormling, I. and Ø. Johnsen. 1992. Effects of the parental
environ-ment on full-sib families of Pinus sylvestris. Can. J. For. Res. 22:88--100.
Gutterman, Y., T.H. Thomas and W. Heydecker. 1975. Effect on prog-eny of applying different day length and hormone treatments to parent plants of Lactuca scariola. Physiol. Plant. 34:30--38. Johnsen, Ø. 1988. After-effects on progenies from orchard clones
moved to non-native environments. In Physiology and Genetics of Reforestation. Proc. of the 10th North American Forest Biology Workshop, Vancouver, B.C. Eds. J. Worrall, J. Loo-Dinkins and D.T. Lester. Faculty of Forestry, Univ. B.C., Vancouver, B.C., Can-ada, pp 1--11.
Johnsen, Ø. 1989a. Phenotypic changes in progenies of northern clones of Picea abies (L.) Karst. grown in a southern seed orchard. I. Frost hardiness in a phytotron experiment. Scan. J. For. Res. 4:317--330.
Johnsen, Ø. 1989b. Phenotypic changes in progenies of northern clones of Picea abies (L.) Karst. grown in a southern seed orchard. II. Seasonal growth rhythm and height in field trials. Scan. J. For. Res. 4:331--341.
Johnsen, Ø. 1989c. Freeze-testing young Picea abies plants. A meth-odological study. Scan. J. For. Res. 4:351--367.
Johnsen, Ø. and G. Østreng. 1994. Effects of plus-tree selection and seed orchard environment on progenies of Picea abies. Can. J. For. Res. 24:32--38.
Johnsen, Ø., J. Dietrichson and G. Skaret. 1989. Phenotypic changes in progenies of northern clones of Picea abies (L.) Karst. grown in a southern seed orchard. III. Climatic damage in a progeny trial. Scan. J. For. Res. 4:343--350.
Johnsen, Ø., G. Haug, O.G. Dæhlen, B.S. Grønstad and A.T. Rogn-stad. 1994a. Effects of heat treatment, timing of heat treatment, and gibberellin A4/7 on flowering in potted Picea abies grafts. Scan. J.
For. Res. 9:333--340.
Johnsen, Ø, O.G. Dæhlen, G. Haug, B.S. Grønstad and A.T. Rognstad. 1994b. Female cone abortion and full seed production in an indoor seed orchard with potted grafts of Picea abies. Scand. J. For. Res. 9:329--332.
Kohmann, K. and Ø. Johnsen. 1994. The timing of bud-set in seedlings of Picea abies from seed crops of a cool versus a warm spring and summer. Silvae Genet. 43:328--332.
Lindgren, D. and R.-P. Wei. 1994. Effects of maternal environment on mortality and growth in young Pinus sylvestris field trials. Tree Physiol. 14:323--327.
Norrell, L., G. Eriksson, I. Ekberg and I. Dormling. 1986. Inheritance of autumn frost hardiness in Pinus sylvestris L. seedlings. Theor. Appl. Genet. 72:440--448.
Philipson, J.J. 1990. Prospects for enhancing flowering of conifers and broadleaves of potential silvicultural importance in Britain. For-estry 63:223--240.
Ross, S.D., A.M. Eastham and R.C. Bower. 1986. Potential for con-tainer seed orchards. In Conifer Tree Seed in the Inland Mountain West. Symposium Proceedings. USDA For. Serv. Gen. Tech. Rep. INT-203, pp 180--186.
Skrøppa, T. 1988. The seed weight did not affect first year’s bud-set in
Picea abies. Scand. J. For. Res. 3:437--439.
Skrøppa, T. 1994. Growth rhythm and hardiness of Picea abies proge-nies of high altitude parents from seeds produced at low elevation. Silvae Genet. 43:95--100.
Skrøppa, T. and Ø. Johnsen. 1994. The genetic response of plant population to a changing environment: the case for non-Mendelian processes. In NATO ASI Series, Vol. I20: Biodiversity, Temperate Ecosystems, and Global Change. Eds. T.J.B. Boyle and C.E.B. Boyle. Springer-Verlag, Berlin, Heidelberg, pp 183--199.
