Black spruce (Picea mariana [Mill.] B.S.P.) seeds were collected from a 64 year-old black spruce stand near Huntsville, Ontario, Canada (45˚26'49.54''N, 79˚14'35.15''W, elevation 281.9 m). The stand was composed of 70% black spruce, 20% balsam fir (Abies balsamea (L.) Mill., and 10% larch Larix laricina (Du Roi) K. Koch) in a lowland riparian zone near Lake Fish. The soil of the stand was organic layers overlain on mineral with rich nutrient contents. The moisture regime was generally very moist to wet. Daily average air temperatures range from −10.3˚C (January) to 18.7˚C (July). Mean annual precipitation is 1105.1 mm (832.2 mm rain and 338 cm snow fall). About 5000 (about 0.1 pound) seeds were collected from 30 trees at least 50 m apart from each other. The seeds were germinated in the greenhouse. 480 germinants were chosen randomly and transplanted into pots (15 cm tall, 13 cm top diameter). The growing medium was a mixture of peat moss and vermiculite (1/1 v/v).

The treatments consisted of two CO₂ concentrations (370 vs. 720 µmol∙mol⁻¹), two photoperiod regimes (photo- period at seed origin: PS; photoperiod at 10˚ north of seed origin: PNM) and two levels of nitrogen supply (30 vs. 300 µmol∙mol⁻¹ N).). The photoperiods for PS were 15 and 10 h for active growing and cold hardening phase, respectively and the respective photoperiods for the PNM treatment were 17 and 8 h. The 15-hour photoperiod represents the actual median photoperiod during the growing season at the seed origin and the 17-hour is the corresponding median photoperiod 10˚ north from the seed origin (Figure 1). The 10 and 8 hour photoperiods represent the median photoperiods during cold hardening phase for the above two locations. 60 µmol∙mol⁻¹ P

and 150 µmol∙mol⁻¹ K were used in both N treatments. N, P and K were formulated using ammonium nitrate (NH₄NO₃), super phosphate (P₂O₅), and muriate of potash (K₂O). The high N concentration and the concentrations of P and K are the optimal levels for the growth of black spruce seedlings (Landis 1989). The experiment was carried out in four greenhouses with identical design features and control devices. Each of the two CO₂ treatments was randomly assigned to two greenhouses. Since the natural day length was shorter than the required photoperiods (started in September), the natural photoperiod was extended using high pressure sodium lamps which gave about 600 µmol∙m⁻¹∙s⁻¹ flux density of photosynthetically active radiation at the canopy level.

The [CO₂], photoperiod, temperature and humidity were automatically controlled and monitored using an Argus control system (Argus Control Systems Ltd, Vancouver, BC, Canada). The same temperature regime (25˚C - 26˚C/16˚C - 17˚C day/night temperature) was used for both PS and PNM treatments because according to the theory of climate envelope the temperature and moisture regimes define the climate envelope and would remain the same after the climate envelope shifts to a new location [2] [3] . The photoperiods in all the greenhouses were set to the value for the PNM and the shorter photoperiod for the PS was achieved by shading the seedlings at both ends of the day. The photoperiod treatments were started with the values for the active growing phase (15 hours for 75 days for PS and 17 hours for 95 days for PNM) and followed by the corresponding values for cold hardening (10 hours for 50 days for PS and 8 hours for 30 days for PNM). The photoperiod and temperature regimes (Figure 1) were determined according to the Astronomical Applications Department (2009) [27] and Hijmans et al. (2005) [28] . The seedlings were fertilized once a week. The water content of the growing medium was maintained above 30% as determined using an HH2 Moisture Meter (Delta-T Devices, Cambridge, U.K.). The seedlings were watered to the dripping point when the volumetric water content of the growing medium declined to 30% [29] .

After 125-days of treatments (Growth Cycle I), height, root collar diameter, terminal bud size, biomass, chlorophyll fluorescence and cold hardiness were measured on seedlings randomly selected from each treatment combination (30 seedlings for height, root collar diameter and terminal bud size, 3 for biomass, 10 for chlorophyll fluorescence, and 15 for cold hardiness). Stem volume (V) was calculated from height (H) and RCD (D) according to van den Driessche (1992): V = (πD²/4)H/3. Biomass was determined separately for shoots and roots after oven-drying at 75˚C for 48 hours in order to calculate root mass ratio (RMR = root mass/total seedling mass).

Chlorophyll fluorescence was measured on the terminal shoot using an FMS 2 pulse modulated chlorophyll fluorometer following a 30 min dark adaptation (Hansatech Instruments Ltd, Norfolk, England). The maximum quantum efficiency of Photosystem II was derived from the initial (Fo) and maximum (Fm) fluorescence (Fv/Fm = (Fm-Fo)/Fm). The length and width of the terminal bud were determined using an AMS-MV2 high-resolution digital imaging system (Advanced Microscopy Group, Mill Creek, WA, USA, 2008) and Micron Imaging Software (USB2) 1.08 (Westover Scientific, Inc., Mill Creek, WA, USA, 2008).

The cold hardiness was assessed according to Colombo et al. [30] [31] . 2 - 3 cm long shoot tips were cut and rinsed with distilled water. Three tips were used for each of the following test temperatures: control (no freezing), −5˚C, −15˚C, −30˚C and −60˚C. Each tip was placed in a separate jar filled with 100 ml distilled water and incubated at room temperature (20˚C - 25˚C) for 24 hours. The electrical conductivity of the solution was measured (referred to as ECC) using an Accumet AR 20 electrical conductivity meter (Fisher Scientific, Ottawa, Canada). Each test temperature was achieved using a programmable freezer (Foster Refrigeration Ltd, King’s Lynn Norfolk, U. K.) by reducing the temperature at a rate of 5˚C/h. One hour after the testing temperature was reached, the temperature was increased back to the room temperature at a rate of 5˚C/h and the electric conductivity of the solution was again measured (referred to as ECF). Subsequently the jars were placed in an oven at 80˚C for two hours. After cooling to the room temperature overnight, the electric conductivity of the solution was measured again (referred to as ECK). The injury index (I_i), which is inversely related to the cold hardiness, was calculated according to Colombo et al. [30] (1984) and Colombo [31] (2003) as follows:

At the completion of all the measurements, all remaining seedlings were cold-stored (−4˚C) for three months for the second growth cycle. The second growth cycle was aimed to examine the treatment effects on bud phenology and seedling survival and growth in the next growing season. The seedlings were exposed to the same growth conditions as in Growth Cycle I. The seedlings in the PS treatment were exposed to short photoperiod (10 hours) for three weeks and then to the long photoperiod (15 hours) for six weeks while the seedlings in the PNM treatment were first exposed to the corresponding short photoperiod (8 hours) for one week and then to the corresponding long photoperiod (17 hours) for eight weeks. The earlier start of the long photoperiod in the PNM treatment was dictated by the faster rate of change in photoperiod at the higher latitude (Figure 1). The bud phenology was observed daily. The total number of days in the long photoperiod regime to bud burst (DBBL) were recorded. Height and RCD were measured again at the end of the second growth cycle. The relative growth rates of height (RGR_h), RCD (RGR_d) and volume (RGR_v) were determined by dividing the growth increments in the second growth cycle by the corresponding measurements at the end of the first growth cycle.

The experiment was a split-plot design with CO₂ treatment as the main plot and factorial combinations of photoperiod and nutrient supply as the split plot. The data were examined graphically for normality (probability plots for residuals) and homogeneity of variance (scatter plots). Both assumptions for the analysis of variance (ANOVA) were satisfied. When ANOVA showed a significant interaction (P ≤ 0.05), the Least Significant Difference (LSD) multiple comparisons were conducted. Since all the seedlings in the low N treatment died in the 2^nd growth cycle, the nitrogen treatment was not included in the analysis for growth cycle II. All the analyses were conducted using the GenStat statistics package, 12th Edition (VSN international Ltd, Hemel Hempstead, UK, 2009).

N supply significantly affected the responses of seedling height, RCD, biomass and biomass allocation to roots to photoperiod and [CO₂] (N × P, N × C and N × P × C in Table 1). While the CO₂ elevation and PNM both increased height and RCD growth, the increases were greater in the high than low N supply (Figure 2). In fact, the increases in RCD were not statistically significant in the low N treatment (Figure 2(c) and Figure 2(d)). The PNM also resulted in a greater increase in seedling biomass in the high than in the low N supply (Figure 3(b)).

The CO₂ elevation decreased the root mass ratio in the low N supply in both photoperiod treatments but increased RMR under the high N supply and PS (Figure 3(a)). While the PNM generally decreased RMR, there were significant interactions with CO₂ and N: the photoperiod impact was greater in the low than the high N treatment under the ambient [CO₂] but the trend was the opposite under the elevated [CO₂], i.e., the impact was greater in the high than the low N treatment (Figure 3(a)).

All the seedlings in the low N treatment died in the second growth cycle and thus N supply was not considered in the second growth cycle. While the positive effect of PNM on height growth remained significant (30.47 ± 1.40 vs. 26.49 ± 1.03 cm, Table 2) in the second growth cycle, the effects of PNM and [CO₂] on RCD became insignificant (Table 2). The relative growth rates of height, stem volume and RCD, however, showed opposite trends to those of height and RCD growth, i.e., PNM reduced the RGR for height (0.44 ± 0.01 for PNM vs. 0.59 ± 0.01 cm∙cm⁻¹ for PS), stem volume (3.23 ± 0.15 vs. 6.39 ± 0.22 cm³∙cm⁻³) and RCD (Figure 3(c)). The CO₂ elevation also significantly reduced the RGR of RCD under the PNM but did not affect it under the PS (Table 2 and Figure 3(c)). Photoperiod also had no significant effects on height increment (Hi) or RCD increment (RCDi) in the second growth cycle. No treatment had a significant effect on total seedling biomass in the second growth cycle (Table 2).

Buds were much longer and wider under the PS than PNM photoperiod regime in the high N treatment, but photoperiod had no significant effect on bud size in the low N treatment (Table 1 and Figure 4). The elevated [CO₂] also did not significantly affect bud size (Table 1).

In the first growth cycle, the maximum quantum efficiency of PSII (F_v/F_m) was greater in the high than low N treatment in both photoperiod treatments, and the PNM significantly increased F_v/F_m only at the high N (P × N interaction in Table 1, Figure 5(a)). In the following growth cycle, the PNM significantly increased F_v/F_m under the elevated [CO₂], and the CO₂ elevation increased F_v/F_m in both photoperiod treatments (C × P interaction in Table 2, Figure 5(b)).

The freezing injury was generally much greater in the low than high N treatment at all the freezing temperatures (−5˚C to −60˚C, Table 1). For example, the injury index at −5˚C was 4.20 ± 0.34 (mean ± SEM) and 2.87 ± 0.20 for the low and high N respectively, and 30.91 ± 0.81 and 15.15 ± 0.89 for the low and high N respectively at −60˚C. N supply also significantly influenced the effects of [CO₂] and photoperiod on freezing injuries at the two modest freezing temperatures, −15˚C and −30˚C (Table 1). While the injury index was generally greater in the PNM than in the PS, the difference was greater in the high N treatment at −15˚C but trend was the reverse at −30˚C (Figure 6(a) & Figure 6(c)). The CO₂ elevation reduced the freezing injury in the low but not in the high N treatment (Figure 6(b)). The CO₂ elevation increased freezing injury in both photoperiod treatments but the PNM increased freezing injury only under the ambient [CO₂] (Figure 6(d)).

Seedling mortality was greater in the PNM than the PS (Table 2; 41.46% ± 4.81% vs. 4.17% ± 0.41%). None of the treatments had a significant effect on the timing of bud burst (DBBL in Table 2).