Materials and methods

Seeds were collected in natural populations of P. cembroides and P. orizabensis in 2014 and 2015 (Table 1). In the case of P. cembroides, the trees were selected because they are geographically close to the other species and at high elevation and thus could be compared. In each provenance, 15 healthy trees with good crown conformation were chosen; the trees were at least 50 m apart. The number of cones collected varied from 10 to 15 per individual tree. The cones from each tree were packed in paper bags labeled with the name of the species, provenance and date. The strobili were transported to the laboratory of the graduate program in Forest Sciences of the Colegio de Postgraduados in Texcoco, state of Mexico, to be processed and the seed analyzed.

Because we observed that each tree produces seed with seed coat of only one color and that producers obtain higher prices with seeds that are not black, we decided to test this factor as well as the species and their geographic origin. The number of seeds of some species-provenance-testa color combinations was small, and for this reason, in the tests a smaller sample size than suggested by the ISTA (^{International Seed Testing Association, 1999}) was used. For the same reason and because we did not find trees that produced certain colors of testa in some provenances, we formed 18 lots of seed depending on the amount available per species, provenance and testa color on two collection dates (Table 2). Each lot is a group of seeds composed of several trees that had seed of a certain test color in each location of a certain year. Differences in germinative capacity (CG), viability, peak value (VP), days to peak value (DVP), days to reach 50 % of the germinative capacity (GERM50), testa thickness and imbibition were determined. In both species three colors were identified (black, brown and bi-color). Only for P. cembroides there were not sufficient bi-color seeds for the tests, and we considered only the other two colors for tests with this species.

The physiological tests were conducted with seed collected in two consecutive years. The 2014 seed was stored at humidity below 8 % in a cold chamber at -6 °C. No significant differences between the collection years were detected in germination tests, and thus, this factor was not considered in later analyses.

Seed moisture content

Moisture content was determined in two seed samples that had initial weight of 3 to 4 g. These samples were placed in aluminum boxes with covers and put into a drying oven Riossa^® at 130 °C for 1 h (^{ISTA, 1999}). The seed was weighed on an analytical balance Edam^® PW124, with a precision of 0.1 mg. Moisture content (MC) was calculated with the following formula (^{Kolotelo et al., 2001}; ^{Bonner et al., 1994}):

Testa thickness

Testa thickness was measured in four replications of 25 seeds per lot. The seeds were broken with a die press C-Clamp (Truper, model 17666 PTR-2). Each seed was placed vertically in the press with the micropyle pointing downward and moderate pressure was exerted until the testa fractured longitudinally. The larger piece of the testa was measured. Measurement of thickness was done at the opposite ends of the testa at a distance of 5 mm above the micropyle near the middle of the testa with a digital vernier (Mitutoyo, model CD-6CS). The value reported was the average thickness for the lot.

Seed imbibition

The seed imbibition curve of the 18 lots was constructed by monitoring seed weight gain. Four replications of 15 seeds per lot were placed in 100 mL distilled water at room temperature in precipitation beakers. Init ial weight was recorded and monitored every 2 h for 24 h and later every 4 h for additional 20h. Weight was quantified on an Edam^® PW124 analytical balance. The water used for each sample was changed every 12 h. The percentage of absorption obtained was accumulated seed weight per unit of time, relative to initial seed weight.

Seed viability

Four replications of 25 seeds from each lot were used. The testa was eliminated, and the embryos and megagametophytes were soaked for 12 h in distilled water. The megagametophyte was cut longitudinally with a #4 scalpel to saturate the embryo in a 1 % solution of 2, 3, 5 tetrazolium chloride (TZ) in glass Petri dishes. The Petri dishes with the samples were placed for 24 h at room temperature in darkness in aluminized bags that did not allow passage of light. The seeds were classified as viable when the megagametophyte and embryo dyed intense red (^{Kolotelo et al., 2001}).

Germination

Before the germination test, the seeds were disinfected in a solution (0.25 %) of commercial sodium hypochlorite (Cloralex^®) by immersion for 5 min plus washing with abundant running water. The testa was scarified with 3 % hydrogen peroxide (v:v) for 2 min and the seed was soaked in distilled water for 12 h before the germination test. The seeds were placed on a cotton cloth in 20 x 27 x 12 cm transparent rigid plastic boxes with hermetic covers. Plastic separators were installed between treatments and 75 mL distilled water was added. They were then sprayed with a fungicide solution of 0.2 % Captán. The boxes with seeds were placed in a chamber with daytime temperature of 24 to 26 °C and night temperature of 22 to 24 °C.

The germinated seeds were recorded every day for 35 d. When the root was as long as the seed, the seed was considered germinated. The variables were germinative capacity (CG), peak value (VP), days to reach peak value (DVP) and time to reach 50 % total germination (GERM50), calculated with the following formulas:

CG (%) = (germinated seeds/total seeds)^*100

VP = MAX Σ germination/days (^{Kolotelo et al., 2001})

DVP (day) = test day on which accumulated germination divided by the number of days lapsed was the highest (^{Kolotelo et al., 2001}).

GERM50 (days) = number of days to reach 50 % of the germinative capacity.

Experimental design and statistical analysis

In the germination test, we used a completely randomized design with 20 replications of 10 seeds per treatment (Table 2). Prior to the statistical analyses, the data on germinative capacity and viability were standardized because it is information expressed in percentage. The percentage data were transformed with the arcsine function of the square root of p (θ =arcoseno √p, where θ is the value of germinative capacity or viability). This transformation is applied to data between 0 and 1 to improve their normal distribution (^{Sokal and Rohlf, 1981}). All the statistical analyses were performed with the MIXED procedure (^{SAS Institute, 2003}). None of the variables were different between years, and this factor was not considered in the statistical models used. Thus, some combinations color-species or provenancespecies had more repetitions.

The models used for the analysis to test differences between species and provenances were:

where Y_ijkl is the observed value of the l^th lot of a certain testa color of the k^th provenance of the j^th species in the i^th replication, µ is the general mean, R_i is the random effect of the i^th replication, E_j is the fixed effect of the j^th species, RE_ij is the random effect of the interaction of the i^th replication with the j^th species, P_k(j) is the fixed effect of the k^th provenance within the j^th species, RP_ijk is the random effect of the interaction of the i^th replication with the kth provenance within the j^th species, and E_ijkl is the experimental error.

To test differences among testa colors, the analysis was performed for each species with the model:

where Y_ijk is the observed value of the k^th lot of a certain provenance of the j^th seed testa color in the i^th replication, µ is the general mean, C_j is the fixed effect of the j^th seed testa color, RC_ij is the random effect of the interaction of the i^th replication with the j^th testa color, and E_ijk is the experimental error.

Moisture content and testa thickness

There were no significant differences between species or among provenances or testa colors. The moisture content of P. cembroides was 0.33 % higher than in P. orizabensis (Table 3). The mean values of seed moisture content with different testa colors varied from 7.24 % to 9.64 %. Seeds of the genus Pinus are orthodox, since they tolerate low moisture content (5-8 %) and remain viable over long periods of storage in controlled temperature with suitable packaging (^{Bonner et al., 1994}).

The differences in seed testa thickness between species and among provenances and testa colors were significant (p≤0.01). The testa of P. cembroides seeds was thicker, and this may have influenced the lower absorption of water by making imbibition difficult (Table 3). Testa of black seeds of both species was thinner, and so they probably absorbed water more easily. ^{Farjon and Styles (1997)} report a greater range of testa thickness for P. orizabensis, 0.7 to 1.1 mm, than for P. cembroides, 0.6 to 1.0 mm.

Seed imbibition

The seeds rapidly absorbed water. After only two hours, P. cembroides seed absorption percentage was 10 % and 12.7 % for P. orizabensis seeds. Later, the rate of absorption decreased but was higher in P. orizabensis seeds. Water absorption of P. cembroides seeds was 25 % at 24 h and at 16 h for P. orizabensis. At the end of the test (44 h), P. cembroides and P. orizabensis had absorbed 32.4 % and 37.6 %, respectively (Figure 1). The highest absorption at the highest rate of the P. orizabensis seeds may be due to the thickness of its testa, which was 0.1 mm thinner than that of P. cembroides.

Figure 1 Accumulated imbibition of Pinus cembroides and P. orizabensis seeds. 

Initial moisture content may have an influence in the seed’s imbibition period (^{Mápula et al., 2008}). In our study, it was 7.8 % for P. cembroides and 7.4 % for P. orizabensis; therefore, absorption was greater in P. orizabensis whose seed was less moist. Soaking in running water for 24 to 48 h is adequate to promote total imbibition of the seeds, and this period varies among conifer species. Seeds of some species can undergo damage if soaking is prolonged (^{Landis et al., 2014}). In Mexican nurseries, the seeds are submerged for 12 or up to 18 h before sowing. Our results indicate that the period of immersion should increase for these seeds, and care was taken so that overexposure to immersion did not have adverse effects on the seeds, such as appearance of pathogens and the lack of oxygenation (^{Landis et al., 2014}). Germination of P. edulis Engelman seeds improved when they were soaked in running water for 48 h (^{Ronco, 1990}).

Black P. cembroides seeds absorbed water faster: 25 % in 22 h, contrasting with 28 h for brown seeds (Figure 2). Black P. orizabensis seeds reached 25 % in 14 h, while brown and bi-colored seeds took 16 and 20 h, respectively (Figure 3).

Figure 2 Accumulated imbibition of Pinus cembroides seeds with testa of different colors. 

Figure 3 Accumulated imbibition of Pinus orizabensis seeds with testa of different colors. 

Seed viability

Differences in seed viability were not significant (p>0.05) between species or among provenances and testa colors. Viability was high in seeds of both species: 93.5 % in P. cembroides and 90.8 % in P. orizabensis (Table 4). The highest values of viability were recorded for P. cembroides collected in Cadereyta (98 %) and the lowest for seeds collected in Colón (89 %). Viability was high for recently collected seeds and for those stored in refrigeration for one year.

Seed germination and vigor

The two species differed significantly in germinative capacity (CG), peak value (VP) and in days to reach 50 % germination (GERM50) but not in days needed to reach peak value (DVP). However, the differences were significant among provenances for the four variables (Table 5).

Germinative capacity of P. cembroides seeds was higher than that of P. orizabensis (Figure 4, Table 6), and germination of P. cembroides seeds was faster, as seen with the fewer number of days required to reach 50 % germination. The peak value shows that P. cembroides seeds are more vigorous, but the days to reach peak value indicate similar germination rate (Table 6).

Figure 4 Germination of Pinus cembroides and P. orizabensis seeds. 

Between 8 and 13 % more of the P. cembroides seed collected in Cadereyta germinated than that collected in Santiago de Anaya and Colón (Figure 5). However, germination rate was higher in seeds collected in Colón, 1.4 d and 3.4 d fewer to reach 50 % germination, compared with seed collected in the other two sites. The peak value shows that seeds from Colón and Cadereyta is more vigorous and seed from Santiago de Anaya is less vigorous.

Figure 5 Germination of Pinus cembroides seed from different provenances. 

P. orizabensis seed germination was more homogeneous and the differences between the extreme values of CG was 5%. Germination of seed from El Carmen was higher (Figure 6) but slower to germinate and with less vigor, around 1 d of difference in reaching 50% of total germination. Altzayanca seed exhibited greater vigor, based on peak value, but Tepeyahualco seed reach peak value more rapidly.

Figure 6 Germination of Pinus orizabensis seed from different provenances. 

The germinative capacity of P. cembroides seed was higher in our study than the value reported for seed of the same species collected in the Ejido La Amapola in San Luis Potosí, where it was 81% (^{Sandoval et al., 2000}). The germinative capacity of P. orizabensis seed in our study was also superior to the same species (48 %) collected in Altzayanca, Tlaxcala (^{Sánchez et al., 2005}); these authors report that germination initiated 30 d after the experiment was set up in agrolite in a greenhouse, unlike our study where it began on the sixth day.

In both pines there were significant differences in germinative capacity among the different testa colors (Table 7). Brown P. cembroides seeds had 5.5% more germination than black seeds, and 8.4% more with brown P. orizabensis seeds. In this species, germination of bicolor and black testa seeds was similar (Table 8; Figure 7). However, only with P. cembroides seed were there differences in days to reach 50 % germination and in number of days to peak value (Table 7). More seeds with brown testa germinated, but P. cembroides seeds with black testa germinated more rapidly, with fewer days to 50 % germination and to peak value. This did not occur with P. orizabensis (Table 8; Figure 8).

Figure 7 Germination curve of Pinus cembroides seeds with different testa color. 

Figure 8 Germination curve of Pinus orizabensis seeds of different testa color. 

^{Martínez et al. (1987)} observed that P. johannis seeds with different testa color were not significantly different in viability, but germination of mottled seed was 36.5 %. This value is a third of that obtained with brown or mottled P. cembroides and P. orizabensis seeds (more than 90 % in both species). The same authors indicate that the seed with lightcolored or black testa had a germination of 84 and 79 %. For the study of P. cembroides and P. orizabensis seed, that with black test had the lowest germinative capacity, 87 and 81 %, respectively. No differences in germination among the different testa colors were reported with P. sylvestris L., and each tree produces testa of different colors (^{Kaliniewicz et al., 2013}).

P. cembroides seed testa is thicker and has a lower rate of water absorption, and contrary to what was expected, more seeds germinated and at a faster rate than P. orizabensis seeds. A prediction could be that the seeds that absorb water at a faster rate, such as those of P. orizabensis, germinate at a faster rate, but it is possible that some physiological mechanism is involved in rapid germination of P. cembroides, and this physiological response could be an adaptive opportunist trait, confirming the existence of morphological and adaptive differences that deserve distinction and management as different taxa, particularly in repopulation that is carried out in the country.

The results of viability and germination indicate the absence of latency in the seed of both species despite the thick testa, which is composed mainly of dead cells and acts as a protective covering for the live inner tissues (^{Kolotelo et al., 2001}). The seeds presented only quiescence and germinate when the environmental conditions are suitable (^{Baskin and Baskin, 2004}).