Responses of Norway spruce seedlings to different ...

478

ARTICLE

Can. J. For. Res. Downloaded from www.nrcresearchpress.com by Metla Finnish Forest Research Inst. on 03/15/16 For personal use only.

Responses of Norway spruce seedlings to different night interruption treatments in autumn Johanna Riikonen and Juha Lappi

Abstract: Photoperiodic lighting can be used in late summer to prevent height growth cessation and terminal bud formation in nurseries growing forest tree species in Nordic countries. To create guidelines for using the method in container nurseries growing Norway spruce (Picea abies (L.) Karst.) and to test the use of light-emitting diode (LED) technology, we exposed first-year, nursery grown seedlings to the following night interruption (NI) treatments from 10 July 2014 onwards (00:00–03:00): (i) no lighting, (ii) 1 min lighting at intervals of 30 min, (iii) 1 min lighting at intervals of 15 min, and (iv) 3 h continuous lighting. Light intensities (LI) of 10, 25, and 70 ␮mol photosynthetically active radiation (PAR) m−2·s−1 were used. Growth, formation of terminal buds, and winter damage of the seedlings were measured. All NI treatments prevented growth cessation at LI of 25 and 70 ␮mol PAR·m−2·s−1, but the intermittent treatments were less effective at a LI of 10 ␮mol PAR·m−2·s−1. The treatments of duration longer than 1 min at intervals of 30 min did not provide any additional increase in shoot growth but predisposed the seedlings to frost injury during autumn and winter. Both seed origins used in this experiment responded similarly to the NI treatments. Key words: container seedling, freezing damage, growth cessation, photoperiodic lighting, Picea abies. Résumé : On peut utiliser l’éclairage photopériodique a` la fin de l'été pour prévenir l’arrêt de la croissance en hauteur et la formation du bourgeon terminal chez les espèces d’arbre forestier cultivées en pépinière dans les pays nordiques. Dans le but d’élaborer des directives pour l’utilisation de cette méthode dans les pépinières de plants en récipient d’épicéa commun (Picea abies (L.) Karst.) et pour tester la technologie DEL, nous avons exposé des semis d’un an cultivés en pépinière aux traitements d’interruption nocturne (IN) a` compter du 10 juillet 2014 (00:00–03:00) : (i) aucun éclairage, (ii) 1 min d’éclairage a` intervalle de 30 min, (iii) 1 min d’éclairage a` intervalle de 15 min et (iv) 3 h d’éclairage continu. Des intensités lumineuses (IL) de 10, 25 et 70 ␮mol de rayonnement photosynthétiquement actif (RPA) m−2·s−1 ont été utilisées. Tous les traitements d’IN ont prévenu l’arrêt de la croissance avec une IL de 25 et 70 ␮mol de RPA m−2·s−1, mais les traitements intermittents étaient moins efficaces avec une IL de 10 ␮mol de RPA m−2·s−1. Les traitements d’une durée de plus de 1 min a` intervalle de 30 min n’ont pas entraîné d’augmentation additionnelle de la croissance des pousses mais ils ont prédisposé les semis aux dommages causés par le gel durant l’automne et l’hiver. Les deux sources de graines ont réagi de façon similaire aux traitements d’IN. [Traduit par la Rédaction] Mots-clés : semis en récipient, dommages dus au gel, arrêt de la croissance, éclairage photopériodique, Picea abies.

Introduction Plants use light as an energy source for photosynthesis and as a key environmental signal that regulates their growth, development, and growth rhythm. Growth cessation and the bud set of northern conifer species are regulated by photoperiod, which is determined by length of the dark period (Dormling et al. 1968; Junttila and Kaurin 1985). Seedlings from northern origins have a shorter critical night length (the shortest dark period that induces bud setting in at least half the plants of a given population) than seedlings from southern origins: they cease their growth earlier in autumn (Vaartaja 1954, 1959). In nurseries in the Nordic countries, forest tree seedlings are generally grown in natural daylight, and artificial light is used only for photoperiod control. Night interruption (NI) treatment is used to prevent premature growth cessation and bud set in seedlings that are sown early in spring when the night is still too long for maintaining growth. In late summer, NI treatment can be applied if seedlings of different latitudinal origins or altitudes are grown simultaneously, if multiple crops are grown in a single growing season, or if there is a need to prevent the cessation of

stem elongation until adequate height growth is achieved (Landis et al. 1992). Photoperiod length is perceived by plants through phytochrome photoreceptors, which exist in a biologically inactive red light (peak absorption at 660 nm) absorbing form and a biologically active far-red light absorbing form (peak absorption at 735 nm) (Pierik and de Wit 2014). Phytochrome is triggered by very low light intensity. In the studies conducted on conifer species cultivated in North America such as Picea glauca (Moench) Voss, Picea engelmannii Parry ex Engelm., and Tsuga mertensiana (Bong.) Carrière (Arnott 1984; Tinus and McDonald 1979), the light intensity required for activating the phytochrome system in early spring was found to be 8–16 ␮mol PAR·m−2·s−1 (see Landis et al. (1992) and references therein), and the lighted period should cover at least 3% of the dark hours (Landis et al. 1992). Similar protocols have also been successfully used in Finnish nurseries for Norway spruce (Picea abies (L.) Karst.) seedlings that are sown in early spring. The use of photoperiodic control in late summer is less well known. There is some indication that the light intensity and duration that is required for keeping Norway spruce seedlings

Received 17 September 2015. Accepted 8 January 2016. J. Riikonen and J. Lappi. Natural Resources Institute Finland, 77600 Suonenjoki, Finland. Corresponding author: Johanna Riikonen (email: johanna.riikonen@luke.fi). Can. J. For. Res. 46: 478–484 (2016) dx.doi.org/10.1139/cjfr-2015-0355

Published at www.nrcresearchpress.com/cjfr on 11 January 2016.


Riikonen and Lappi

actively growing in a late growing season may be higher than that required in spring (Simak 1975). In a preliminary test (unpublished results) that was completed prior to conducting the experiment reported here, we exposed first-year nursery-grown Norway spruce seedlings to a NI treatment (either 1 or 2 min lighting at 30 min intervals from 00:00 to 03:00, 25 ␮mol PAR·m−2·s−1) starting from the beginning of August. The NI treatments were unable to prevent growth cessation and bud set but induced remarkable lammas growth in autumn. The failure may have been caused by an inadequate duration or intensity of the light exposures during the night breaks, or alternatively, the seedlings may have become sensitive to photoperiod earlier than could have been expected based on the critical night length estimated for Norway spruce seedlings originating from Central Finland (Junttila and Nilsen 1993). In container nurseries, the light sources providing the NI treatments are often mounted on a watering boom that moves back and forth above the canopy during the dark period. For technical reasons, the low intensity and duration of light received by the seedlings limits use of the NI method. The weights of the greenhouse lamps that are attached to the watering booms also cause problems for the functionality of the system. Light-emitting diodes (LEDs) are an alternative to traditional greenhouse lamps due to their lower weight, configurable light spectral composition, longevity, and energy efficiency. It is estimated that the costs of LED lighting will continue to decrease, which would facilitate the use of LEDs in agricultural applications (Pinho et al. 2013). The LED technique has recently also been successfully applied for providing supplemental light (Apostol et al. 2015) or as a sole source of radiation (Riikonen et al. 2016) for conifer seedlings. The aim of the current study was to investigate the application of the NI treatment provided by LED lamps to prevent growth cessation and terminal bud set in late summer in container nurseries and to answer the following questions: (i) What is the required duration and frequency of NI treatment for preventing growth cessation and terminal bud formation in late summer? (ii) What is the required light intensity? (iii) Do Norway spruce seedlings originating from 62°N and 64°N latitudes respond to the NI treatments in a similar manner? (iv) Do different NI treatments predispose seedlings to frost injury?

Materials and methods Plant material The experiment was carried out in a greenhouse at the Finnish Forest Research Institute in Suonenjoki, Finland (62°39=N, 27°03=E, 142 m above sea level (a.s.l.)). Two latitudinal origins of Norway spruce were used: 62°N (stand collected and source identified) and 64°N (registered seed orchard no. 366, Paronen). The seeds were sown on 24–26 March 2014 (64°N) and 23–25 April 2014 (62°N) into hard-walled plastic containers (Plantek PL81, BCC, Iso-Vimma, Finland; 81 seedlings per tray, cell volume 85 cm3, 549 seedlings·m−2), which were filled with base-fertilised sphagnum peat. Until June 2014, the seedlings originating from 64°N were grown at the nursery of UPM-Kymmene Corporation in Joroinen and then transported to the research nursery in Suonenjoki. The different sowing dates of the two seed origins are explained by the different growth schedules at the nurseries. The seedlings were grown in a greenhouse according to Finnish nursery practice for one-year-old seedlings using standard fertilisation and irrigation procedures (Juntunen and Rikala 2001). Experimental design The experiment consisted of 12 blocks of trays, which were randomly divided into the three NI treatments and a control treatment; each treatment was replicated three times. Each block consisted of three trays (Fig. 1): one tray each of local (62°N) and northern (64°N) origins were arranged side by side and another

479

Fig. 1. Diagram showing the positions of trays representing seed origins 62°N and 64°N within a block and the light intensities (LI) within the block. LI treatments: LI10, 10 ␮mol PAR·m−2·s−1; LI25, 25 ␮mol PAR·m–2·s–1; LI70, 70 ␮mol PAR·m–2·s–1. Stars denote the seedling used for the growth measurements.

tray of local origin was placed beside them. The blocks were located in a greenhouse, and the distance between the blocks was 4 m. The seedlings received natural daylight, and in addition, each experimental unit (excluding controls) was exposed to a NI treatment, provided by a LED tube (60 W; Kasvua Oy, Hämeenlinna, Finland) that was attached to a frame surrounding each block. Spectral distribution of light under the LED tubes was measured using a spectroradiometer (FieldSpec Pro FR, Analytical Spectral Devices, Boulder, Colorado, USA), as described in Riikonen et al. (2016). The LED tube emitted 70% red (600–700 nm), 25% blue (400–500 nm), and 5% far red (700–800 nm) from the total photon flux density between 400 and 800 nm. The following NI treatments were applied from 10 July to 30 September 2014 and were given nightly from 00:00 to 03:00 using timers: (i) no lighting (control), (ii) 1 min lighting at intervals of 30 min (NI1/30), (iii) 1 min lighting at intervals of 15 min (NI1/15), and (iv) 3 h continuous lighting (NI3h). The selection of the NI treatments was based on studies by Tinus and McDonald (1979) and Arnott and Mitchell (1982) who found that the duration of the dark period between NIs should not exceed 30 min. The LED tube was placed above the middle point of the central trays (20 cm above the canopy), and thus, the arrangement of the local trays enabled us to create different NI light intensities within each group (i.e., various distances between the seedlings and the LED tube) (Fig. 1). The light intensity (LI) within different distances from the LED tube was measured in a dark room before the experiment, with a quantum sensor (LI-190SB, LI-COR, Inc., Lincoln, Nebraska, USA). The LI at the sampling points within each NI treatment was 70 ␮mol PAR·m−2·s−1 (LI70) in the central trays and 25 ␮mol PAR·m−2·s−1 (LI25) and 10 ␮mol PAR·m−2·s−1 (LI10) in the side trays (local origin only). The LI measurements made in the greenhouse showed that the light emitted by the LED tubes placed above the adjacent blocks did not overlap. The light treatment thus consisted of two treatment factors, NI and LI. The seedlings were held in the greenhouse until 11 November 2014 when they were transferred to an open compound. From 20 December 2014 onwards, the seedlings over-wintered under snow cover. The data showing the length of the photoperiod and radiation outside and the air temperature and accumulation of temperature sum inside the greenhouse during the growing season is presented in Figs. 2A and 2B. Published by NRC Research Press

480

Can. J. For. Res. Vol. 46, 2016


Fig. 2. (A) Radiation (hourly mean values) and length of photoperiod in Suonenjoki and (B) average temperature (daily mean values) and accumulation of temperature sum inside the greenhouse in 2014–2015. Date format is as follows: 1.5, 1 May; 1.6, 1 June, etc. NI, night interruption; d.d., degree days.

Measurements Growth measurements were taken from five seedlings per LI treatment (on 10 seedlings per tray; Fig. 1). The height and stem diameter were measured at the beginning of the experiment in July. At the end of the experiment, the height, stem diameter, and dry mass (DM) of roots, stem, and needles (65 °C, 72 h) were determined. The stem volume was calculated using the following model:

共兲

1 d 2 ␲ h 3 2 where d is stem diameter and h is height, thus assuming that the shape is a cone. Using a more accurate form factor would not change the interpretation of the results. The timing of bud set was monitored for all seedlings in the trays, excluding the outer rows. A terminal bud was scored as formed when it had become visible. The number of seedlings with a damaged terminal bud or browning of the needles within each tray was counted on 11 June 2015. Because there was a LI gradient within each tray, it was not possible to use the same LI treatments as for the other variables measured. The data was pooled over each tray — the central trays were exposed to a higher LI (62°N and 64°N) and the side tray was exposed to a lower LI (62°N) (i.e., the light gradient of the side tray was not taken into account) (Fig. 1).

Statistical analyses Because seedlings growing under the control treatment were not exposed to any light treatment, NI and LI factors could not be used as standard statistical treatment factors in the analysis. Dummy (zero-one) variables were thus coded to describe the treatment structure. In the model, “Intercept” represented the control treatment, and the light treatments were compared with the control. Next, the parameter “Treatment” was coded for describing how the light treatments deviated from the control treatment, and the corresponding dummy variable was zero for control and one elsewhere. The reference for comparing the LI treatments was LI70, thus additional dummies were generated for LI25 and LI10. The reference for NI treatments was NI3h, thus additional dummies were generated for NI1/15 and NI1/30. The corresponding interaction dummies were generated for NI × LI. A dummy for seed origin 64°N was created, and this dummy was multiplied with the Intercept dummy, Treatment dummy, and NI dummies to create models where the seedlings of different seed origins behaved differently in the control treatment and in the NI treatments. As there was only one LI treatment (LI70) for seed origin 64°N, the seed origin × LI interactions could not be estimated. In the statistical analyses, initial height, stem diameter, and stem volume were used as covariates. Height, stem diameter, and stem volume growth were used to estimate the treatment effects. Note that the use of height, stem diameter, and volume as dependent variables at the end of the experiment would have caused high trivial correlations with the covariates. The residual error of the model was decomposed into two components, the tray effect and the seedling effect, and a mixed linear model was used (SPSS 22.0 for Windows; Chicago, Illinois, USA). The mixed model was used to take into account that seedlings within the same tray were correlated. The correlations of the random tray effects and residual errors across the different dependent variables were also estimated using the multivariate linear mixed models. Technically, these multivariate models were formulated as univariate models using dummy variables. It is computationally demanding to estimate all correlations simultaneously, thus the correlations were estimated pairwise. The maximum likelihood method was first used for the estimations to enable the use of the likelihood ratio test for testing the fixed effects of the models. First, a full model (including the covariates, interactions of covariates and seed origins, NI treatments, interactions of NI treatments and seed origins, LI treatments, interactions of LI and NI treatments, and interaction of Treatment and seed origin) was generated. Nonsignificant treatment effect dummies, excluding Treatment, were then dropped from the model. The likelihood ratio test was used to test whether this model deviated from the full model. Note that because the different parameter estimates were correlated in the full model, it can easily be the case that not all nonsignificant effects can be dropped from the model. If the restricted model deviated from the full model, we tried to include such explanatory variables in the model that the model would agree with the full model. After finding the final form of the model, restricted maximum likelihood (REML) was used to estimate the model parameters. REML is considered to be better estimation method for small data sets, but REML log likelihood cannot be used to compare models with different fixed effects. To simplify the estimation procedure, the covariates and the interaction of seed origin and covariates were kept in the model even if they were not significant. Note that significance of the Intercept is not of interest. When Treatment was significant, it indicates an effect of light treatment on the variable measured. The proportion of seedlings with damaged terminal bud or browning of the needles was analysed using a generalized linear mixed model (see Stroup (2013)). The tray effects were assumed to be random effects. The proportions were assumed to have a binomial distribution. A logit link was used to link the probability of Published by NRC Research Press

Riikonen and Lappi

481

Table 1. The linear mixed model estimates and statistics. Variable

Parameter

Estimate

Standard error

Significance

Height growth (141 mm)

Treatment NI1/30 LI10 NI1/15 LI10 Treatment LI10 Treatment Treatment Treatment Treatment Treatment Treatment

78 –38 –29 –124.3 53.3 104.4 81.1 78.3 0.06 138.5 –0.13

10 12 12 21.1 15.2 52.1 36.1 98.4 0.08 39.4 0.02

control (Table 5; Fig. 4). No difference was noted between the two LI treatments (P = 0.594), and thus, LI was dropped from the final model.

Discussion According to Junttila and Nilsen (1993), seedlings from northern ecotypes cease growing in mid-August under local light regimes. According to our preliminary test (unpublished data), however, seedlings originating from 62°N had already become insensitive to the NI treatments at the beginning of August. In this experiment, the NI treatments prevented bud formation and height growth cessation in Norway spruce seedlings when they were started in mid-July. Published by NRC Research Press

Riikonen and Lappi

483

Table 3. Estimated variances of the tray and seedling effects, standard errors of the variance estimates (SE), and the square roots of the variance estimates (SD) derived from linear mixed models.


Tray effect

Seedling effect

Variable

Variance

SE

SD

Variance

SE

SD

Height growth Root DM Needle DM Stem DM Total DM Diameter growth Volume growth Root to shoot ratio

312 1161 6689 3256 23378 0.02 4507 0.002

110 439 2577 1232 9164 0.01 1465 0.001

18 34 82 57 153 0.14 67 0.040

1225 6288 40199 18933 147478 0.08 16314 0.003

97 496 3163 1489 11601 0.01 1283 0.001

35 79 201 138 384 0.28 128 0.059

Note: The SD can be compared with the means of the variables and parameters shown in Table 2. DM, dry mass.

Table 4. Correlation between the tray effects and between the seedling effects on different variables. Variables

Tray effect

Seedling effect

Height growth vs. diameter growth Height growth vs. volume growth Height growth vs. needle DM Diameter growth vs. stem DM Volume growth vs. diameter growth Volume growth vs. root DM Needle DM vs. stem DM Needle DM vs. root DM Diameter vs. root DM Shoot DM vs. root DM

0.274 0.999 1.000 0.905 0.977 0.266 0.922 0.424 0.428 0.677

0.473 0.737 0.582 0.793 0.889 0.615 0.849 0.629 0.685 0.435

Fig. 4. Proportion of seedlings with damaged terminal bud and (or) browning of needles (mean ± 1 standard error plotted; n = 3) in Norway spruce seedlings originating from 62°N and 64°N and grown with or without night interruption (NI) treatments. The NI treatments were applied nightly from 00:00 to 03:00 from 10 July 2014 to 30 September 2014: C, no lighting; NI1/30, 1 min lighting at intervals of 30 min, NI1/15, 1 min lighting at intervals of 15 min; NI3h, 3 h continuous lighting. The damage was measured on 11 June 2015. Values for the probability of damage according to the generalised linear mixed model (presented in Table 5) are shown in parentheses above each bar.

Note: DM, dry mass.

Table 5. The parameters of the linear predictor of the logit of the probability in the generalised linear mixed model describing the proportion of damaged seedlings. Parameter

Estimate

SE

Significance

NI1/15 NI3h Seed origin 64°N

0.84 1.41 –1.13

0.39 0.39 0.30

0.040 0.001 0.001

Note: The mean proportion of damaged treated seedlings was 19%. The estimated variance of the tray effect was 0.544 and standard error (SE) was 0.234. NI1/15, 1 min lighting at intervals of 15 min; NI3h, 3 h continuous lighting.

All NI treatments used in this experiment prevented growth cessation and terminal bud formation, and even the minimum frequency of 1 min lighting every 30 min was adequate. Consequently, the seedlings grown under the NI treatments were taller with greater needle and stem DM compared with control seedlings. The shoot growth under the NI treatments occurred at the expense of root DM growth, which is in accordance with a study on P. engelmannii provenances (Arnott 1979). Increasing the duration or frequency of the NI treatment from 1 min lighting at 30 min intervals did not cause a further increase in height growth, providing that the LI was 25 ␮mol PAR·m−2·s−1 or above; no additional benefit from increasing the duration of NI treatment was observed. The LI needed for photoperiod control is very low compared with that needed for driving photosynthesis. According to an earlier experiment where seedlings of P. abies were illuminated nightly for 1 h, an LI of about 20 ␮mol PAR·m−2·s−1 (1000 lx) was required to prevent height growth cessation in the autumn (Simak 1975). In our experiment, the LI treatment of 10 ␮mol PAR·m−2·s−1 maintained height growth in a similar manner as the LI treatments of higher light intensities when seedlings were lighted continuously for 3 h, but it was less effective in the intermittent NI treatments (NI1/30 and NI1/15). This is in accordance with Young and Hanover (1977), who found that low LI can be compensated by longer NI duration in Picea pungens Engelm. Thus,

on one hand, the LI during intermittent lighting should be 25 ␮mol PAR·m−2·s−1, but when continuous lighting is used for 3 h, the LI of 10 ␮mol PAR·m−2·s−1 is adequate for keeping Norway spruce seedlings actively growing. On the other hand, increasing the LI from 25 to 70 ␮mol PAR·m−2·s−1 provided no additional benefit to the height growth of the seedlings. All the other variables associated with height growth, stem volume, and stem DM were also affected by light treatments. The effect of light treatment on these variables was not as significant as on height growth. The interactions NI1/15 × LI10 and NI1/30 × LI10 were not needed in the model. The diameter growth was unaffected by light treatment, and this made the effect of light treatment weaker with respect to stem volume and stem DM. Light treatments decreased root DM. When comparing the models for height growth and root DM, it was noted that root DM increment cannot be analysed in a similar way as height growth was analysed. Although the initial height was also a significant covariate for root DM, it is a less accurate covariate as is initial root DM. It is logical that treatments would differ in a similar way with respect to height growth and root DM; however, when the explanatory variables used for height growth (i.e., Treatment, NI1/30 × LI10, and NI1/30 × LI10) were used for root DM, then NI1/30 × LI10 Published by NRC Research Press


484

and NI1/30 × LI10 were not significant, and the Akaike's information criterion was clearly worse than with the model where LI10 was a predictor via Treatment. The light treatments increased height growth and decreased root DM in a similar fashion and, consequently, did not have any effect on total DM. It is known that the use of photoperiod lighting late in the growing season will predispose seedling shoots to injury from possible early autumn frosts and that these treatments should be terminated in time for the seedlings to have sufficient time to develop frost hardiness before the first frosts in autumn (Arnott 1984). In this experiment, the NI treatments were continued until the end of the September to determine whether the growth could be maintained, although the length of the natural photoperiod, light level, and temperature will drop in autumn. Indeed, the seedlings that received one of the NI treatments continued growing until the end of the experiment. Apparently, shoot lignification was very incomplete and predisposed the shoots to frost damage in autumn. Although the average height, DM accumulation, and the visual appearance of the seedlings among the NI treatments were similar, the proportion of damaged buds in the consequent spring was higher the longer the duration of NI. We assume that main reason for seedling damage was the frost episodes in mid-October (first frosts on 17 October; Fig. 2B), but other factors may have been involved such as drought or sugar depletion during the overwintering. The seedlings of both seed origins behaved similarly, although the seedling damage was more abundant in seed origin 62°N. The seedlings originating from 62°N were sown a month later than seedlings originating from 64°N, and thus, they were in an earlier phase of their development during the experiment. This may explain the higher proportion of damaged seedlings in spring in seed origin 62°N, in addition to its more southern seed origin. Altogether, the effect of seed origin is confounded with the effect of sowing time. Why the seedlings that received the NI3h treatment had the highest vulnerability is unknown. The metabolic activity of these seedlings may have been higher than in the seedlings grown under the NI treatments of shorter durations, predisposing them to frost damage and to other stress factors. The vulnerability of the seedlings to damages in the terminal buds during autumn and winter was not dependent on the LI in NI treatment.

Conclusions According to our results, 1 min lighting at 30 min intervals for 3 h was adequate to prevent bud formation and height growth cessation in Norway spruce seedlings in autumn when the LI was 25 ␮mol PAR·m−2·s−1. A higher duration or frequency of the NI treatment or LI did not provide any additional increase in shoot growth; on the contrary, the seedlings that received longer durations of NI treatment were more susceptible to frost damage. The LED technique was successfully used as a source of photoperiodic light.

Acknowledgements This study was funded by the European Social Fund (ESF, project S11890) and the Finnish Forest Research Institute (project 3554). We thank the staff of the research nursery in the Natural Re-

Can. J. For. Res. Vol. 46, 2016

sources Institute Finland, Suonenjoki unit, for maintaining the experiment, and UPM-Kymmene Corporation, Joroinen nursery, for producing the seedlings originating from 64°N. We are grateful to Dr. Risto Rikala and Dr. Jaana Luoranen for their valuable comments on the manuscript.

References Apostol, K.G., Dumroese, R.K., Pinto, J.R., and Davis, A.S. 2015. Response of conifer species from three latitudinal populations to light spectra generated by light-emitting diodes and high-pressure sodium lamps. Can. J. For. Res. 45(12): 1711–1719. doi:10.1139/cjfr-2015-0106. Arnott, J.T. 1979. Effect of light intensity during extended photoperiod on growth of amabilis fir, mountain hemlock, and white and Engelmann spruce seedlings. Can. J. For. Res. 9(1): 82–89. doi:10.1139/x79-014. Arnott, J.T., and Mitchell, A. 1982. Influence of extended photoperiod on growth of white and Engelmann spruce seedlings in coastal British Columbia nurseries. In Proceedings, Canadian Containerized Tree Seedling Symposium, Toronto, ON, 14–16 September 1981. Edited by Scarratt, J.B., Glerum, C., and Plexman, C.A. Ontario Ministry of Natural Resources. COJFRC Sym. Proc. O-P10: 139–152. Arnott, T.A. 1984. Photoperiod control of container seedlings. In Proceedings, Western Forest Nursery Council and Intermountain Nurseryman’s Association. Coeur d’Arlene, ID, 14-16 August 1984. Edited by Landis, T.D. USDA Forest Service, Ogden, Utah, General Technical Report INT-185, pp. 9–13. Dormling, I., Gustafsson, Å., and von Wettstein, D. 1968. The experimental control of the life cycle in Picea abies (L.) Karst. I. Some basic experiments on the vegetative cycle. Silvae Genet. 17: 44–64. Junttila, O., and Kaurin, Å. 1985. Climatic control of apical growth cessation in latitudinal ecotypes of Salix pentandra L. In Plant production in the North. Edited by Kaurin, Å., and Nilsen, J. Norwegian University Press, Tromsø, Norway. pp. 83–91. Junttila, O., and Nilsen, J. 1993. Growth and development of northern forest trees as affected by temperature and light. In Forest development in cold climates. Edited by Alden, J., Mastrantonio, J.L., and Ødum, S. Plenum Press, New York. pp. 43–57. Juntunen, M.L., and Rikala, R. 2001. Fertilization practice in Finnish forest nurseries from the standpoint of environmental impact. New For. 21: 141–158. doi:10.1023/A:1011837800185. Landis, T.D., Tinus, R.W., McDonald, S.E., and Barnett, J.P. 1992. Atmospheric environment. In The container tree nursery manual. USDA Forest Service, Washington DC, Agriculture Handbook 674. Vol. 3. Pierik, R., and de Wit, M. 2014. Shade avoidance: phytochrome signalling and other aboveground neighbor detection cues. J. Exp. Bot. 65: 2815–2824. doi: 10.1093/jxb/ert389. PMID:24323503. Pinho, P., Hytönen, T., Rantanen, M., Elomaa, P., and Halonen, L. 2013. Dynamic control of supplemental lighting intensity in a greenhouse environment. Lighting Res. Technol. 45: 295–304. doi:10.1177/1477153512444064. Riikonen, J., Kettunen, N., Gritsevich, M., Hakala, T., Särkkä, L., and Tahvonen, R. 2016. Growth and development of Norway spruce and Scots pine seedlings under different light spectra. Env. Exp. Bot. 121: 112–120. doi: 10.1016/j.envexpbot.2015.06.006. Simak, M. 1975. Kort nattbelysning av skogsplantor I plastväxhus ger batter odlingsmaterial. Summary: Intermittent light treatment of forest plants in a plastic greenhouse produces better plant material. Sveriges Skogsvårdsförbunds Tidskrift, 4: 373–381. Stroup, W.W. 2013. Generalized linear mixed models. CRC Press, Boca Raton, Florida. Tinus, R.W., and McDonald, S.E. 1979. How to grow tree seedlings in containers in greenhouse. Gen. Tech. Rep. RM-60. Fort Collins, CO: USDA Forest Service, Rocky Mountain Forest and Range Experiment station. 256 p. Vaartaja, O. 1954. Photoperiodic ecotypes of trees. Can. J. Bot. 32(3): 392–399. doi:10.1139/b54-036. Vaartaja, O. 1959. Evidence of photoperiodic ecotypes in trees. Ecol. Mono. 29: 91–111. doi:10.2307/1942199. Young, E., and Hanover, J.W. 1977. Effects of quality, intensity, and duration of light breaks during a long night on dormancy in blue spruce (Picea pungens Engelm.) seedlings. Plant Physiol. 60: 271–273. doi:10.1104/pp.60.2.271. PMID: 16660073.

Published by NRC Research Press