To understand differences in germination responses among circumboreal populations of V. opulus, we conducted a series of experiments in a controlled (laboratory) environment as well as in a common garden. We examined (i) embryo size and seed size and (ii) specific temperature requirements for root and shoot emergences among a subset of 12 populations from Asia, North America and Europe. For all 12 populations, we determined the timing of root and shoot emergences, and of embryo growth, in relation to a sequence of simulated seasonal temperatures. Lastly, emergence phenology of roots and shoots was observed in a common garden in Sweden for 11 of the 12 populations.
The fruit is a one-seeded 10- to 15-mm-diameter drupe, with soft pulp and a flattened stone; the fully ripened fruit is red. Fruits were collected from all populations in autumn 2004 (Table ). Distances between populations were: 1–65 km in Sweden, 175–710 km in Japan, 1000 km in Canada and the USA, 5200–6200 km between Europe and North America, 7600–8200 km between Europe and Asia, and 9600–10 300 km between Asia and North America (Google Inc. 2009
). Experiments began 1–22 days after collection and shipment to laboratories. Pulp was removed from the stones (hereafter referred to as seeds) and seeds were dried at room conditions for studies.
Collection locations and dates for seeds from three varieties of V. opulus.
General laboratory procedures
Laboratory work on Asian and North American seeds was performed at Middle Tennessee State University in Tennessee, USA (MTSU), and that on European seeds at Linköping University in Sweden (LiU). A preliminary study showed that the timing of emergence and the percentages of emergence for roots and shoots from seeds collected in Tokarp (Sweden) and in Sapporo (Japan) did not differ between laboratories.
Incubators (MTSU: I-36LL, Percival Scientific, Perry, IA, USA; LiU: Rumed Cooled 3122, Rubarth Apparatebau, Laatzen, Germany) were set at alternating 12/12 h (day/night) temperatures of 15/5, 20/10, 25/15 and 30/15 (or 30/20) °C. A constant temperature of 5 °C was achieved in another incubator (MTSU) or in a cold room (LiU; room: Ki-PANEL, Huurre Sweden AB, Sweden; temperature control: ADU 200, Styrprojektering AB, Sweden). We used the same set of temperature regimes among all populations to simulate seasons: 5 °C, winter; 15/5 °C, early spring and late autumn; 20/10 °C, late spring and early autumn; and 25/15 and 30/15 (or 30/20) °C, summer. With the exception of photoperiod, light conditions were similar in incubators and the cold room [described by Walck and Hidayati (2007)
for MTSU and by Karlsson et al. (2005)
for LiU]. The photoperiod was 14 h (MTSU) or 12 h (LiU), which apparently had no effect given that root and shoot emergences were similar between laboratories in the preliminary study (see above).
Since germination in V. opulus has two distinct phases, temporally separated, we use the term ‘rootling’ for seedlings in the first phase of germination (i.e. a seed with only an emerged radicle and no shoot) to distinguish it from seedlings in the second phase (i.e. a seed with a root and shoot). To examine root emergence, seeds were placed on two filter papers moistened with distilled water in 10-cm-diameter Petri dishes and incubated in light. Dishes were wrapped with plastic film and placed in polyethylene plastic bags to reduce water loss during the experiment, and additional water was added if needed. Roots were determined to have emerged when the radicle protruded ~1 mm beyond the seed. To study shoot emergence, a rootling was removed from the dish and planted in a 7-cm-deep plastic container, filled with 2–3 cm of moist sand, and covered with plastic film; only the root and lower portion of the seed were covered with sand. Shoots were judged to have emerged when at least one cotyledon was completely exposed from the seed coat. Seeds/rootlings were either placed in one dish/container or divided equally among two to three dishes/containers.
Lengths of 20–25 embryos were obtained each time by excising them from seeds using a razor blade and measuring them with a micrometer under a dissecting microscope. At the start of the study, seed and embryo lengths were determined for fresh seeds 24 h after they were placed on moist filter paper at room temperature (~22 °C). Measurements for fresh seeds and embryos were obtained for Mombetsu, Sapporo, Sendai (Japan), Moncton (Canada), Ithaca (USA) and Tokarp (Sweden) populations; the embryo : seed (E : S) ratios were calculated from these measurements.
Root and shoot emergences were evaluated fortnightly. Percentages of root emergence are based on number of viable seeds, and those of shoot emergence on the number of seeds in which a radicle had emerged. Seeds from which radicles did not protrude were checked to determine whether the embryos were white and firm, indicating that they were viable, or if the embryos were brown and soft, indicating that they were nonviable. Throughout all laboratory studies, 70–97 % of seeds maintained their viability. Survivorship of rootlings until the end of all experiments varied from 62 to 93 % among populations, except for 22 and 39 % in the Kaminokuni and Sendai (Japan) populations, respectively.
Specific requirements for root emergence
Twenty-six Mombetsu (Japan), 150 Moncton (Canada) and 50 Tokarp (Sweden) seeds were used for each treatment. Dishes were placed at 5, 15/5, 20/10, 25/15 °C and either 30/15 (MTSU) or 30/20 °C (LiU) for 12 weeks of stratification and were then transferred to (or remained at) each of these temperature regimes for 12 weeks of incubation, e.g. seeds stratified at 5 °C were incubated at 5, 15/5, 20/10, 25/15 °C and either 30/15 or 30/20 °C. Seeds were checked for an emerged radicle fortnightly, and if present, counted and removed from the dish. At the end of the experiment, seeds from which radicles did not protrude were checked for viability.
Specific requirements for shoot emergence
Several hundred seeds from the Moncton (Canada), E Stocklycke and Tokarp (Sweden) populations were placed at 25/15 °C to allow radicle emergence. After many seeds had radicles, the experiment was set up using 15, 33 or 40 seeds, respectively, per treatment. They were cold stratified at 5 °C for 0, 12 and 24 weeks (Canadian population) or 0, 6, 12 and 24 weeks (Swedish populations). After each cold period, seeds were incubated at 15/5, 20/10 and 25/15 °C for 12 weeks. Seeds in which the shoot emerged during stratification were removed from the containers before incubation and subtracted from the total tested. Control seeds were kept at 5 °C for 36 weeks. Rootlings were checked for an emerged shoot fortnightly and, if present, were counted and removed from the dish.
Temperature sequence for root and shoot emergences and for embryo growth
Seeds from all populations were exposed to two conditions. (i) In continuous temperature regimes, seeds were incubated at 5, 15/5, 20/10 and 25/15 °C for 80 weeks. (ii) In annual temperature cycles, which simulated a natural sequence of seasonal temperatures, seeds were given a winter (5 °C) regime progressing to a summer (25/15 °C) regime (cold (C) temperatures → warm (W) temperatures) or a summer regime progressing to a winter regime (W → C) for two annual cycles equalling 80 weeks. The number of seeds used in this study was: 45 for Sapporo (Japan); 132 for Kaminokuni (Japan); 100 for Sweden (except for Tokarp); 150 for North America, Mombetsu and Sendai (Japan), and Tokarp (Sweden). Rootlings were removed from dishes, planted in containers, and the cycle continued.
Embryo growth was examined for Mombetsu, Sapporo (Japan), Moncton (Canada) and Tokarp (Sweden) populations. In each of the continuous temperature regimes, embryos were measured following 12 and 36 weeks of incubation at 5, 15/5, 20/10 and 25/15 °C for Japanese populations, and following 6, 12, 24 and 36 weeks of incubation for the Canadian and Swedish populations. In the annual temperature cycles, embryos from these four populations were measured each time dishes or containers were moved to a new temperature regime (for up to 60 weeks). Embryo length was measured only in ungerminated seeds.
Phenology of root and shoot emergences in a common garden
With the exception of Mombetsu (Japan), all populations were used in this study. Japanese and North American seeds were sent to LiU 1–2 days after arriving at MTSU. Fifty seeds were sown on top of soil on 13 November 2004 (except for Sendai, sown on 1 January 2005) in each of 2–5 (2 L) pots per population. The soil used was found locally (Ledberg) and sterilized for 20 min at 120 °C before being placed in pots containing a non-woven glass-fibre sheet in the bottom. The pots were placed at Ledberg, Sweden (58°26′N, 15°28′E), close to the Tokarp population. A 0.15-m-deep hole was dug in the ground and filled with 2- to 6-mm-diameter ceramic clay pellets (AB Svenska, Leca, Sweden) before pots were added. The pots were buried to a depth such that the surface of the soil was at the same level as the surface of the surroundings. They were protected from direct wind, but otherwise subjected to natural weather conditions. Animals were excluded by a 20-mm-mesh net. Temperature (without protection from sunshine) was measured every hour (TinytagPlus, Intab, Sweden) at the soil surface in an additional pot and precipitation was measured daily, ~15 m away from the pots. On 7 October 2005, 1–2 pots from each population were examined for rootlings, and after recording their number they were replanted. Shoot emergence was examined fortnightly throughout the study, except when the pots were covered with snow. At the end of the study (31 August 2007), the number of remaining seeds (with and without a radicle) that had not produced a shoot during the study were counted. Percentages of root and shoot emergences were calculated based on the number of seeds sown at the start of the study.
For embryo/seed size data and root/shoot emergence data, we determined 95 % confidence intervals. This approach, as opposed to testing specific null hypotheses, allows the presentation of the magnitude of effect size as well as the precision of the estimate for this magnitude (Cumming and Finch 2005
; Nakagawa and Cuthill 2007
). Confidence intervals for emergence data were calculated using the exact method for binomial parameters. A principal components analysis (PCA) was conducted to summarize the complex data for 26 characters (Table ) per studied population from the seasonal sequence experiment. Principal components analysis was applied with the purpose of displaying the relative similarity in characters among the populations (software CANOCO 4.5, ter Braak and Smilauer 2002
). The 26 characters were normalized before the PCA. The length of an arrow indicates relative importance to the solution and the angle of one arrow relative to another shows the relative correlation between them. The same data set was also used to calculate Euclidean distances between populations, i.e. their relative similarity in the seasonal sequence experiment. This Euclidean distance matrix was then compared with a corresponding one on geographic distances by a Mantel test. Pair-wise geographic distances were log(x
+ 1) transformed.
Characters, and their acronyms, used in the Mantel test and PCA.