Nursery stage

The nursery stage was carried out in the experimental greenhouses belonging to the Colegio de Postgraduados’ Postgraduate Forest Sciences program in Montecillo, Texcoco, State of Mexico (9° 29’ N - 98° 54’ W and elevation of 2 240 m). P. rudis was planted during the first two weeks of October 2012. The seed came from the "El Vergel" germplasm bank, operated by the National Forestry Commission network and located in San Pedro Cholula, Puebla. A mixture of pine bark, peat moss, perlite and vermiculite was used as substrate at a ratio of 70:10:10:10 by volume. In addition, Osmocote Plus^® fertilizer (15-9-12) was added with a release time of nine months, in doses of 7 kg·m^-3 of substrate, which is equivalent to 3 750, 967.5 and 2 490 ppm in mass of N, P and K, respectively, taking into account that the density of the substrate was 0.28 g·cm^-3. Individual black containers (tubes) with lateral openings were used, measuring 11.9 cm long, 6 cm in upper diameter, and 4.9 cm in lower diameter, with a volume of 220 mL. The plants grew in a plastic-covered greenhouse for eight months before starting the hardening test. Nursery management of the plant included daily irrigations to field capacity during the first establishment stage (six weeks) and irrigations to saturation every other day during the fast growth stage. Only in the last month of this eight-month preparatory phase were supplementary applications of soluble fertilizer (Peters Professional^®) made to the irrigation water once a week. This is a common practice in the conifer plant production process. Formulation 20-20-20 was used in doses of 0.5 g·L^-1, equivalent to 100, 43 and 83 ppm (m/v) of N, P and K, respectively.

Fertilization treatments in the hardening stage

Hardening started on May 31, 2013 and consisted of using traditional fertilization (control) and increasing combinations of P and K. The control hardening treatment corresponded to plants fertilized with Peters Professional^® (4-25-35 N, P and K, respectively), according to the recommended levels for the hardening phase of conifers (^{Landis, 1989}). Fertilization combinations were prepared according to each treatment shown in Table 1. Treatments 2 and 3 consisted of adding P at two levels: 50 % (P+) and 100 % (P++) with respect to the control. Similarly, for treatments 4 and 5, two levels identified as K+ and K++ were used, corresponding to a 50 and 100 % increase of K with respect to the control. Finally, in treatments 6 and 7, the dose of Peters Professional^® was increased by 50 % (PT+) and 100 % (PT++).

In the treatments, P and K levels were modified with 85 % phosphoric acid (H₃PO₄) and potassium sulfate (K₂SO₄), respectively. Irrigation combined with fertilization was done every 72 h for a period of eight weeks. The experiment was established in a completely randomized design. Each of the seven treatments (Table 1) was made up of 50 plants and replicated four times, resulting in a total of 1 400 plants.

Evaluation of morphometric characteristics

After hardening, 48 plants randomly selected per treatment were evaluated; 12 plants were extracted from the central part of the plots from each replicate to eliminate the possible edge effect. Root collar diameter was measured in each plant and then the substrate was carefully removed with abundant water so as not to lose any plant material of interest. Subsequently, the aerial part was separated from the root with a cut at the height of the root collar; excess moisture was removed with absorbent paper and each part was deposited in a paper bag. The samples were dried in an oven with air circulation at 70 °C for 72 h and then the dry weight of each component was obtained. The aerial part/root ratio (APRR) was calculated from the data.

Evaluation of root growth potential

The root growth potential test consisted of placing a random sample of plants in a favorable environment to promote rapid root growth. These plants grew in a substrate with a 70:30 ratio of bark and perlite for 40 days in 10-L pots, protected under a greenhouse and with daily irrigation. At the time of transplant all white roots were cut to give rise to the formation of fine roots (<2 mm). White roots represent newly formed tissue, whose growth potential is influenced by the plant's ability to adapt to site conditions (^{Ostonen, Lõhmus, Helmisaari, Truu, & Meel, 2007}). The experimental unit consisted of three plants per treatment with four replicates, giving a total of 12 plants per treatment. The arrangement of the pots was random. At 40 days, the plants were removed from the pots and the roots were carefully washed for measurement. The variables obtained were total number of initial roots (NIR), number of roots formed (NRF) and dry weight of new roots (DWNR). New roots were identified by their characteristic white color, as has been done in similar studies (^{Sánchez-Aguilar, Aldrete, Vargas-Hernández, & Ordaz-Chaparro, 2016}).

Nutritional analysis

After hardening, the NPK nutrient concentration was analyzed in the foliage and in the aerial part (four replicates per treatment). Foliar analyses were interpreted using the vector analysis technique proposed by ^{Park, Park, and Bae (2015)}. This technique analyzes, together, the relative changes of the variables plant component mass and nutrient concentration and content (^{Isaac & Kimaro, 2011}; ^{López-López & Alvarado-López, 2010}; ^{Mead, Scott, & Chang, 2010}; ^{Park et al., 2015}; ^{Salifu & Timmer, 2003}). Vectors were plotted with a program based on macros programmed in Excel (http://bit.ly/2Eok8GF). Figure 1 shows a summary of the graphical basis of the vector method used in this study and the explanation of the changes with respect to the control treatment. The main effects were determined according to the graphical interpretation, as has been done in other studies (^{Isaac & Kimaro, 2011}; ^{López-López & Alvarado-López, 2010}; ^{Mead et al., 2010}; ^{Park et al., 2015}; ^{Salifu & Timmer, 2003}).

Figure 1 Vector method in nutritional analysis. Graphical summary of the main changes in the variables (mass of the plant component and concentration and content of nutrients) and their most likely effect. 

Statistical analysis

An analysis of variance (P ≤ 0.05) was performed to compare the means of the morphometric variables, root growth potential and nutrient concentrations among treatments, under a completely randomized model with the use of the SAS version 9.0 program (^{Statistical Analysis System Institute [SAS], 2002}). Separation of means was done with the Tukey test (P ≤ 0.05).

Morphometric characteristics of Pinus rudis

According to Table 2, the morphometric characteristics of P. rudis showed no significant differences among treatments (P > 0.05). In 10 months, P. rudis seedlings reached a root collar diameter of 4.8 to 5.5 mm, aerial dry weight of 4.2 to 4.4 g, root dry weight of 1.1 to 1.2 g and the aerial part/root ratio (APRR) in all treatments was 4.

Morphometric parameters indicate that, despite undergoing different levels of fertilization, biomass proportions did not change. The results suggest that for P. rudis, biologically it is more important to maintain its morphometric balance than to take advantage of greater nutrient availability, as occurs in other forest species such as Quercus ilex ssp. ballota (Desf.) Samp. (^{Andivia, Fernández, & Vázquez-Piqué, 2011}). The ideal APRR varies between 2 and 3 (^{Jackson et al., 2012}; ^{Villar, 2003}), but in this study P. rudis showed a value of 4. It is possible that because it is a species that grows in sites with a dry spring season and a cold winter, the priority of energy investment in root biomass is lower compared to other species (^{Magaña, Torres, Rodríguez, Aguirre, & Fierros, 2008}).

The increase in P, K and FS during the eight weeks of hardening had no significant effect on morphometry (P > 0.05). In this respect, in Pinus engelmannii Carr., fertilization hardening combined with exposure to the elements stimulated root growth and promoted a greater concentration of nutrients in the foliage (^{García-Pérez et al., 2015}); however, the net advantages in the field have not yet been demonstrated due to the lack of research and knowledge about pine species in Mexico. Although Pinus montezumae Lamb. showed a favorable response to fertilization hardening in the field, this benefit was only observed in northern exposures with higher water availability and lower temperatures (^{Robles, Rodríguez, & Villanueva, 2017}). In studies with Pinus resinosa Ait., it was observed that, although nutrient concentrations during the hardening phase did not differ between fertilized and unfertilized plants, the former responded faster in growth after a period of drought (^{Miller & Timmer, 1994}). These papers highlight the importance of continuing with hardening studies and field evaluation.

Root growth potential

Most treatments showed the NIR and NRF to be similar to those of the control (Table 3). Only the K+ and PT++ treatments showed significant difference (P ≤ 0.05) in NIR from each other, as did P+ and PT++ in the NRF (P ≤ 0.05). Regarding the DWNR, no significant differences were observed among treatments; however, when the P dose increased by 50 %, the seedlings increased their number of white roots by 25 % compared to the control.

A notable result was the highest NIR obtained due to the 50 % increase in the dose of K (K+). The treatment with 50 % more phosphorus (P+) generated higher NRF with an average of 201; in contrast, a 100 % increase in soluble fertilizer (PT++) resulted in the lowest NRF. This is because the roots acquire P mainly by diffusion and the plants tend to form more fine roots to ensure their supply (^{Schlesinger & Bernhardt, 2013}); therefore, if the substrate provides enough P, fewer fine roots are generated. The results show a bell trend with an optimum of 259 mg P per plant that is satisfied with the P+ treatment.

Root growth potential is a reliable index of successful plant establishment in the field (^{Grossnickle, 2005}) because it provides data on the production and growth of new roots during the rooting of seedlings in an optimal environment (Hasse, 2007).

Nutritional analysis

Concentrations of N, P and K in the foliage, after the hardening process, showed positive and negative changes with respect to the control, but remained within the expected values for conifers (^{Landis, 1989}). With the PT++ treatment, the highest concentration of N was obtained with a proportion of 17.78 mg·g^-1, while the lowest concentration of N was recorded in the P++ treatment with a value of 13.69 mg·g^-1. The highest concentration of P was obtained with the P++ and PT++ treatments with values ​​of 2.96 and 2.74 mg·g^-1, respectively. In the case of K, the highest concentration was obtained with the PT++ treatment with a value of 5.57 mg·g^-1. The variation in K concentrations (4.71 to 5.57 mg·g^-1) is within the values recorded for other pine species such as Pinus devoniana Lindl. (^{Rueda et al., 2010}). N and P concentrations are those expected for conifers under fertilization experiments and higher than those expected without fertilization; that is, values higher than 10 mg·g^-1 N and 1.2 mg·g^-1 P (^{Nambiar & Fife, 1991}). The concentration of nutrients in the foliage and in the aerial part showed the same trend (data not shown).

Vector analysis

Since the results showed a similar graphical dispersion of nutrient concentration for the aerial and foliage biomass variables, it was decided to use a single variable for the analysis. The results shown correspond to aerial biomass, a variable that represents an important store of reserves in seedlings (^{Grossnickle & MacDonald, 2018}). Graphical analysis of the treatments showed two predominant directions in the vectors in relation to aerial biomass. The first direction indicates an increase in nutrient concentration and content, but no change in the biomass category (diagonal lines) (^{Park et al., 2015}). This orientation is defined by vector D in Figure 1. The second direction is related to the decrease in nutrient concentration and content without change in the biomass category. In this case, the orientation is defined by the vector H in Figure 1, which indicates an antagonistic effect (^{Isaac & Kimaro, 2011}).

Figure 2 shows the vector trend for relative aerial biomass (change in aerial biomass with respect to the control) of this study. The PT++ treatment promoted higher luxury consumption for all three nutrients. The vectors are identified as N-PT++ and K-PT++, and luxury consumption is denoted by the greater displacement of each vector towards the upper right corner in each graph in Figure 2; i.e., an increase in concentration and net mass, but without proportional change in relative biomass (compare with vector D in Figure 1).

The P++ treatment (100 % higher dose than the traditional one) stimulated luxury consumption of P, since it was effective in increasing the content by almost 30 % (Figure 2); however, it reduced the N concentration (Table 4; Figure 2). The direction of the K concentration vectors in relation to the treatments is similar to those of N and P, but of shorter length, which is due to a lower response.

Table 4 Concentration of N, P and K in the foliage of Pinus rudis after fertilization hardening treatments. 

Treatment	Nitrogen (mg·g-1)	Phosphorus (mg·g-1)	Potassium (mg·g-1)
P+	14.60 ± 1.53 bc	2.42 ± 0.07 b	4.71 ± 0.44 b
P++	13.69 ± 0.84 c	2.96 ± 0.17 a	5.43 ± 1.34 b
K+	14.70 ± 0.60 bc	2.24 ± 0.08 b	5.30 ± 0.60 b
K++	16.81 ± 1.96 ab	2.19 ± 0.09 b	5.32 ± 0.25 b
PT+	16.90 ± 1.38 ab	2.39 ± 0.14 b	4.85 ± 0.52 b
PT++	17.78 ± 1.10 a	2.76 ± 0.25 a	5.57 ± 1.02 a
Control	15.95 ± 1.43 abc	2.34 ± 0.14 b	5.20 ± 1.40 b

Treatments: Control = Peters Professional^® (4-25-35); P+ and P++ = 50 and 100 % increase in P, respectively, in relation to the control; K+ and K++ = 50 and 100 % increase in K, respectively, in relation to the control; PT+ and PT++ = 50 and 100 % increase in Peters Professional^®, respectively. Average values ± standard error. Means with different letters are statistically different according to Tukey's test (P ≤ 0.05).

The objective of fertilization hardening is to increase the nutrient content in the plant mass. In this regard, the gains in nutrient concentration of this study are considered low, since conifers achieve increases of 50 % compared to the control (^{Dumroese, 2003}), however, what is remarkable is that the relative increases were detectable by the graphical method.

Figure 2 Vector diagram for nutritional analysis in Pinus rudis aerial biomass. Graphs a, b and c show changes in the relative concentration and content of N, P and K. In graph d, the black, red and green line vectors correspond to the P++, PT++ and K++ treatments, respectively. In the names of the vectors, the first letter indicates the nutrient analyzed and the second (after the hyphen) corresponds to the treatment (control = Peters Professional^® [4-25-35]; P+ and P++ = 50 and 100 % increase in P, respectively, in relation to the control; K+ and K++ = 50 and 100 % increase in K, respectively, in relation to the control; PT+ and PT++ = 50 and 100 % increase in Peters Professional^®, respectively).  

The results of this work show two strategies for hardening P. rudis. The first is the additional load of P in the double application of soluble fertilizer in the irrigation (PT++), which is also associated with higher concentrations of foliar N and K. This hardening strategy can be considered when the P. rudis plant is to be established in low fertility soils, such as Cambisols or Leptosols. The second scenario is an additional load of P with the double application of phosphorus-based fertilizer (P++); however, this practice would be more advisable when the plant is destined to sites where P is critical, as is the case of Andosols.