Genetic control of quantitative and qualitative traits of Calycophyllum spruceanum in the Peruvian Amazon

Abstract

Families of 38 month-old Calycophyllum spruceanum planted at three locations in the Peruvian Amazon were analyzed for breeding purpose. Significant differences were detected among families, sites and combined sites for the traits trunk diameter (D), height (H), number of nodes (NN), number of branches (NB) and survival rate (SUR), except for NN at site 2 and SUR at site 3 and combined sites. The correlation of the genotype-environment interaction was simple for D, H, NN, and SUR (0.77-0.89). Heritability was generally highest for D, H, NN, and NB (0.12-0.7), and the genotypic coefficient of variation ranged from 8.6-27.4%. Genetic correlations between these traits within and among sites ranged from 0.35-1.0. The selection of 75 trees from different families, within blocks and locations, suggested the possibility of genetic gains (6.1-29.8%) for D, H, NN, and NB, and that the remaining effective population size can be exploited in future selection cycles.

Keywords:
Genetic correlations; genetic variation; heritability; progeny test; tree breeding

INTRODUCTION

Calycophyllum spruceamun (Benth.) (Rubiaceae) is a heliophytic pioneer tree found in alluvial forests in Brazil, Bolivia, Colombia, Ecuador and Peru, up to 1000 m above sea level (Reynel et al. 2003, Sears et al. 2014, Ramírez 2016, Saldaña et al. 2021). The tree has multiple uses; its timber is exported to various parts of the world in view of the beautiful creamy brown color and high density of the wood, which makes it durable and suitable for construction (Taylor 2005). The very straight, cylindrical bole of adult trees can reach diameters of 50-120 cm and a height of 20-35 m, with a crown in the upper third. A study on this species with seeds from different origins in the Peruvian Amazon reported growth to a height of 1.4-1.6 m within 6 months and to 3.5-4.7 m within one year (Sotelo Montes et al. 2003). In Peru, the response potential of the species in plantations for industrial purposes is considered highly efficient and promising (Orrego and Bustamante 2017). At the national and international level, the demand for C. spruceanum is on the rise, due to the wood properties and low production costs. In recent years, the species has been counted among the first 10 commercial species for saw wood production, with very diverse applications, e.g., as structural timber for construction, furniture and sporting goods (Orrego and Bustamante 2017, Saldaña et al. 2021).

Due to the economic and social relevance of C. spruceanum wood production by small farmers, the genetic variability of wild populations of the species has been the focus of some research based on genetic markers (Russel et al. 1999, Dávila-Lara et al. 2017, Saldaña et al. 2021) and quantitative traits (Sotelo Montes et al. 2003, Sotelo Montes et al. 2006, Sotelo Montes et al. 2007a, Sotelo Montes et al. 2007b, Sotelo Montes et al. 2007c, Sotelo Montes et al. 2008, Boivin-Chabot et al. 2004, Weber and Sotelo Montes 2005, Dawson et al. 2009, Weber et al. 2009, Thauchen et al. 2011). However, information about the genetic variation and control of growth and quality traits of the species is currently limited, since these studies addressed only a minor part of the species distribution.

Due to the increase in deforestation, forest fires and intense selective cutting in the environments of C. spruceanum, as well as the increased demand for its seeds for reforestation (Carmo et al. 2017), strategies of conservation and breeding of the species are urgently needed. Provenance and progeny tests, underlying the determination of genetic variation within and among populations, genetic inheritance (heritability) and phenotypic and genetic correlations between quantitative and qualitative traits have been used for both ex situ conservation and breeding of populations (Sebbenn et al. 2005). These data are fundamental to evaluate whether a germplasm collection strategy efficiently represents and conserves the genetic variation of populations of a species ex situ and for successful selection in breeding programs (Sebbenn et al. 2009a). Provenance and progeny tests can therefore be used for both purposes simultaneously, ex situ conservation and production of somewhat improved seeds, for both commercial and environmental reforestation (Sebbenn et al. 2005, Sebbenn et al. 2009a). To this end, the best trees with desirable traits are selected (Sebbenn et al. 2009a, Canuto et al. 2016). Likewise, testing families from different origins allows a better understanding of the genetic variability in populations of a species and detects the most adaptable genotypes to the environmental conditions of a reforestation site (Menegatti et al. 2016, Revilla-Chávez et al. 2022).

Therefore, this study analyzed the genetic variation and control of quantitative and qualitative traits of 38-month-old C. spruceanum trees in progeny trials in the Peruvian Amazon, to obtain information for the implementation of second-generation seed orchards with high seed quality for forest restoration.

MATERIAL AND METHODS

Study sites and progeny tests

The progeny tests were established at three locations in the Aguaytía river basin, on the eastern slope of the Cordillera, in the department of Ucayali, Peru (between lat 9º 24' 25.87" S, long 75º 54' 38.91" W and lat 7º 57' 36.47" S, long 74º 5' 33.64" W, in an area 100 km E-W and 60 km N-S). The vegetation of the area is mainly tropical premontane forest, with a mean annual temperature of 26 °C (Holdridge 1978). Rainfall increases with increasing elevation: at sites 1, 2 and 3, at average elevations of 150, 200, and 260 m asl, respectively, mean annual precipitation is 1656.3, 1926.45, and 2501.0 mm ha^-1 year^-1, respectively (Hijmans et al. 2005). However, water stress of 1, 4 and 5 months was observed at sites 3, 2 and 1, respectively (Ugarte-Guerra and Domínguez-Torrejón 2010). Data of the progeny tests and characteristics of the three experimental sites are listed in Table 1.

Table 1. Location,
progeny test and edaphoclimatic characteristics of evaluation plots of Callycophyllum spruceanum at three sites in the Aguaytia river basin, Ucayali, Perú

Description of the trials	Site 1	Site 2	Site 3
Number of families	200	200	200
Number of blocks	5	5	5
Number of plants per plot	2	2	2
Elevation (m)€	150	200	260
Rainfall (mm/ha/year)€	1656	1926	2501
Number of months under water stress£	5	4	1
Terrace€	low	medium	high
Soil€	very sandy	sandy loam	clayey
Phosphorus availability (P)	low	low	high
Levels of K, Ca and MG€	medium	medium	high

• ^€
  Hijmans et al. (2005); ^£Ugarte-Guerra and Domínguez-Torrejón (2010); K= Potassium; Ca= Calcium; Mg= Magnesiun

The progeny tests were arranged in a randomized complete block design, with five blocks, 200 open-pollinated families and two plants per plots. The families were derived from seeds collected from 200 seed trees in June 1998, at 13 sites in the Aguaytia river basin: 39 families from the district of Nueva Requena; 5 families from the Neshuya river, at km 49.5 of the countryside highway Carretera Federico Basadre (CFB); 5 families from the Tahuayo river basin, at km 72 CFB; 10 from the Curimaná river basin; 6 from the Aguaytía river basin; 7 from the Yurac-Aguaytía river basin; 21 from Puerto Inca; 6 from Von Humboldt; 43 from Macuya; 50 from San Alejandro; 7 families from along the CFB up to km 72; 6 from along the Curimaná road; and 3 families from the Aguaytía river basin along the CFB to the district of Tournavista (Table 2). Each block comprised a total of 400 plants, 200 plots and 2 plants per plot, at a spacing of 2.5 x 2.5 m, on an area of 2,500 m² per block. Two plant rows around the blocks were considered border rows that were not evaluated and they were separated by a 5-m wide fire strip. At the time of planting, 1600 kg of earthworm humus ha^-1 and 320 kg of rock phosphate ha^-1 (1 kg and 200 g per plant, respectively) were applied to the experimental area. In the second and third years, fertilization was repeated, with 112 kg ha^-1 urea, 316 kg ha^-1 rock phosphate and 160 kg ha^-1 potassium chloride (i.e., 70, 185, and 100 g of each product per plant, respectively). During the first 12 months, plants stunted by weeds, insects and rodents were eliminated.

Table 2
Score used to measure stem shape (SSh) in 38-month-old Calycophyllum spruceanum trees

Class	Terminal but presence	Terminal but dominance	Branch orientation	Branch position	Stem	Score
Stem shape	yes	yes	vertical	different	straight	100
	yes	yes	horizontal	different	straight	98
	yes	yes	vertical	different	crooked	96
	yes	yes	horizontal	different	crooked	94
Branched stem	yes	yes	vertical	different	-	92
	yes	yes	horizontal	different	-	90
	yes	yes	vertical	single plane	-	88
	yes	yes	horizontal	single plane	-	86
	no	yes	vertical	different	-	84
	no	yes	horizontal	different	-	82
	no	yes	vertical	single plane	-	80
	no	yes	horizontal	single plane	-	78
	yes	no	vertical	different	-	76
	yes	no	horizontal	different	-	74
	yes	no	vertical	single plane	-	72
	yes	no	horizontal	single plane	-	70
	no	no	vertical	different	-	68
	no	no	horizontal	different	-	66
	no	no	horizontal	single plane	-	64
	no	no	vertical	single plane	-	62
Inclined stem shape	yes	yes	vertical	different	-	60
	yes	yes	horizontal	different	-	58
	yes	yes	vertical	single plane	-	56
	yes	yes	horizontal	single plane	-	54
	no	yes	vertical	different	-	52
	no	yes	horizontal	different	-	50
	no	no	vertical	single plane	-	48
	no	no	horizontal	single plane	-	46
	yes	no	vertical	different	-	44
	yes	no	horizontal	different	-	42
	yes	no	vertical	single plane	-	40
	yes	no	horizontal	single plane	-	38

Estimates of genetic parameters

The data were arranged in a randomized block design and processed as described below. The REML/BLUP procedure (restricted maximum likelihood/best unbiased linear prediction) by the linear mixed model was used in individual analyses to estimate the variance components and genetic parameters of each trait evaluated at each site, using the SELEGEN software, based on the statistical Model 1 (Resende 2016), $y=Xr+Za+Wp+e$, where $y$, $r$, $a$, $p$ and $e$ are the data, effects of repetition (fixed) added to the overall mean, individual additive genetic effects (random), effects of portion and the error or residual vector (random), respectively. Capital letters represent the incidence matrices for the said effects. For combined site analysis, Model 4 (Resende 2016) was used: $y=Xr+Za+Wp+Ti+e$, where $y$ is the data vector; $r$ the vector of repetition effects (fixed) added to the overall mean; $a$ the vector of individual additive genetic effects (random); $p$ the vector of plot effects (random); $i$ the vector of the effects of the genotype - environment (GE) interaction (random); and $e$ the error or residual vector (random). Capital letters represent the incidence matrices for the said effects. By analysis of variance, the following components of interest were estimated: ${\sigma }_A^2$ = additive genetic variance; ${\sigma }_f^2$= genetic variance among families; ${\sigma }_e^2$= variance due to family × replication interaction effects; ${\sigma }_{e\left(s\right)}^2$= variance due to the interaction between the families within and replications within environments; ${\sigma }_{gxe}^2$= variance of the GE interaction; and = phenotypic variance within family. The genetic correlations between sites (${\mathrm{r}}_{\mathrm{g}\mathrm{e}}$), mean heritability among families ($h_m^2$) and within family ($h_w^2$) for each site were estimated by: $h_m^2=\frac{{\sigma }_f^2}{{\sigma }_f^2+\frac{{\sigma }_e^2}{J}+\frac{{\sigma }_w^2}{\bar{n}J}}$ and $h_w^2=\frac{(1-0.25){\sigma }_A^2}{{\sigma }_w^2}$, and for combined analysis of sites (s) by: $h_{m\left(s\right)}^2=\frac{{\sigma }_f^2}{{\sigma }_f^2+\frac{{\sigma }_{gxe}^2}{S}+\frac{{\sigma }_{e\left(s\right)}^2}{SJ}+\frac{{\sigma }_w^2}{\bar{n}SJ}}$ , and $h_{w\left(s\right)}^2=\frac{\left[\right(1-0.25)/\bar{n}]{\sigma }_A^2}{{\sigma }_{e\left(s\right)}^2+\frac{{\sigma }_w^2}{\bar{n}}}$, where $S$, $J$ and $\bar{n}$ are the number of sites, replication and the harmonic mean of the number of plants per plot. We also calculated the coefficients of environmental variance ($C{V}_e\left(\%\right)=100(\sqrt{{\sigma }_e^2}/x))$, genetic variance among families ($C{V}_g\left(\%\right)=100(\sqrt{{\sigma }_g^2}/x)$), and additive genetic variation among families ($C{V}_{gi}\left(\%\right)=100(\sqrt{{\sigma }_A^2}/x)$, where $x$ is the mean of the trait under study. The relative correlation coefficient was estimated by $C{V}_r=C{V}_g\left(\%\right)/C{V}_e\left(\%\right)$. The genetic ($r_g$) and phenotypic ($r_f$) correlations between pairs of traits were computed as proposed by Resende (2002), taking only traits with genetic variation among families into consideration. The genetic gain by selection among and within families ($G_{aw}$) was estimated for each site for selection intensities of 37.5% of the best families (70:200) and 25% of the best trees within families (5:10), based on the equation, $G_{aw}\left(\%\right)={100(d}_a{h}_m^2+d_w{h}_w^2)/x$), where $d_a$ and $d_w$ are the selection differentials among and within families, respectively. $d_a$ was estimated by $d_a$=$x_{sa}-x$, where $x_{sa}$ is the mean of the selected families and $x$ the population mean for the traits under study. $d_w$ was estimated by $d_w$=$x_{sw}-x$, where $x_{sw}$ is the mean of selected trees within family for the traits under study.

The effective population size ($N_e$) after selection was estimated as $N_e=0.5/\Theta$, where$\Theta$ is the group coancestry (Lindgren et al. 1996), estimated by $\Theta =[mn0.5\left(1+F_p\right)+{m\theta }_{xy}n\left(n-1\right)]/(nm{)}^2$ (Sebbenn et al. 2009b), where m and n are the number of selected families and plants within family, respectively; $F_p$ is the inbreeding coefficient in the parental population; and ${\theta }_{xy}$the coancestry between plants within family, assumed as half-sibs (${\theta }_{xy}$= 0.125).

RESULTS AND DISCUSSION

Significant differences among families at sites and combined sites were detected for D, H, NN, NB and SUR, with exception for NN at site 2, and SUR at site 3 and in the combined analysis. These results indicated genetic variability among families for D, H, NN, NB and SUR and the possibility of breeding by selection of the best families (Table 3). The correlation of the GE interaction (${\mathrm{r}}_{\mathrm{g}\mathrm{e}}$) was positive and high for D, H, NN, and SUR (${\mathrm{r}}_{\mathrm{g}\mathrm{e}}$: 0.768-0.891), indicating that the GE interaction is simple. Little family rank changes for the GE interaction between sites were observed, indicating that selection for these traits can be done efficiently at a single site. In contrast, the ${\mathrm{r}}_{\mathrm{g}\mathrm{e}}$ for NB (0.554) and SF (0.154) indicated a complex type GE interaction, where family rank changes between sites are involved, and that selection for these traits should be performed specifically for each site.

Table 3
Mean and estimates of genetic parameters for trait at site and combined sites in tests with 38-month-old Calycophyllum spruceanum progenies

	Trait	${\mathit{h}}_{\mathit{m}}^2$	${\mathit{h}}_{\mathit{w}}^2$	${\mathit{r}}_{\mathit{a}}$	${\mathit{C}\mathit{V}}_{\mathit{g}\mathit{i}}\mathit{}$ (%)	${\mathit{C}\mathit{V}}_{\mathit{r}}$	$\mathit{G}\mathit{S}$ (%)	${\mathit{r}}_{\mathit{g}\mathit{e}}$
Diameter (D, cm)
Site 1	4.7*	0.269	0.220	0.519	11.9	0.27	18.4
Site 2	8.6*	0.464	0.416	0.681	18.3	0.42	27.7
Site 3	11.2*	0.512	0.383	0.715	14.5	0.46	19.5
Combined	8.16*	0.654	0.329	0.808	15.8	0.40	26.5	0.84
Height (H, m)
Site 1	3.2*	0.203	0.201	0.450	13.7	0.22	13.2
Site 2	6.41*	0.555	0.520	0.745	13.5	0.44	23.9
Site 3	8.52*	0.549	0.502	0.741	14.5	0.49	14.5
Combined	6.4*	0.696	0.429	0.834	16.1	0.50	16.6	0.891
Stem shape (SSh)
Site 1	94.8	0.027	0.008	0.163	1.1	0.07	0.1
Site 2	95.3	0.030	0.012	0.172	0.9	0.08	0.1
Site 3	95.6	0.035	0.012	0.186	0.9	0.08	0.1
Combined	95.2	0.024	0.003	0.154	0.5	0.03	0.1	0.154
N. of nodes (NN)
Site 1	26.9*	0.141	0.121	0.375	8.6	0.18	7.0
Site 2	37.0	0.276	0.238	0.526	9.6	0.28	8.8
Site 3	42.6*	0.300	0.233	0.548	8.8	0.29	6.1
Combined	35.5*	0.473	0.263	0.687	12.0	0.35	25.5	0.85
N. of branches (NB)
Site 1	7.9*	0.379	0.376	0.616	27.4	0.35	29.8
Site 2	13.4*	0.321	0.288	0.567	17.9	0.32	11.7
Site 3	17.6*	0.392	0.372	0.626	11.0	0.31	8.7
Combined	12.9*	0.404	0.197	0.636	13.7	0.29	16.6	0.554
Survival (SUR, %)
Site 1	86.6*	0.196	-	0.443	11.9	0.22	1.7
Site 2	93.5*	0.039	-	0.197	3.2	0.09	0.2
Site 3	96.9	0.023	-	0.151	1.8	0.07	0.1
Combined	92.3	0.292	-	0.540	6.6	0.14	2.4	0.768

• * P< 0.05, with 0.5 degrees of freedom for likelihood ratio test (LRT), χ² deviance chi-square;

  rge

  is the genetic environmental correlation;

  hm2

  and

  hw2

  are the mean heritabilities among and within family, respectively;

  ra

  is the selective accuracy;

  CVgi

  and

  CVr

  are the coefficients of additive genetic and relative variation, respectively.

The mean values of all traits were lowest at site 1 and highest at site 3 (Table 3). After 38 months, the mean survival rate (SUR) among sites (92.3%) indicated a high adaptation of the families to the three studied sites. The mean growth among sites for D (8.16 cm) and H (6.4 m) was greater than the mean reported for the species at other locations in Peru after 39 months (D= 5.86 cm, H= 6.21 m; Sotelo Monte et al. 2006). In contrast, the NN (35.9) and NB (12.9) were lower than the mean reported at other locations in Peru after 18 months (NN= 41.21, NB= 23.49; Sotelo Monte et al. 2003). These results showed that the families studied here have the potential to produce good quality wood (low number of knots in the wood), since the low values of NN and NB indicate wood with less defects (knots).

The traits D, H, NN, and NB generally have higher mean heritability among families ($h_m^2)$, heritability within family ($h_w^2$)$,$ selective accuracy ($r_a$), coefficient of genotypic additive genetic variation ($C{V}_{gi}$), and coefficient of relative variation ($C{V}_r$) than stem shape (SSh) and SUR, suggesting that high genetic gains by selection can be expected for D, H, NN, and NB. The coefficient $h_m^2$ was higher than $h_w^2$ for all traits (except for SUR) in the site and combined site analysis ($h_m^2$: 0.024-0.696, $h_w^2$: 0.003-0.52), indicating that higher genetic gains can be expected from selection among than within families. For SUR, the range of $h_m^2$ was 0.033 to 0.196 among sites (combined= 0.292). The selective accuracy ($r_a)$ was 0.375 to 834 for D, H, NN and NB; 0.443 for SUR at site 1; and 0.54 for the combined sites, which indicated that the phenotype mean is a good predictor of the additive genetic value of families.

Our results showed that the genetic parameters were generally lower for D, H, and NN and higher for SUR at site 1 than at the other two locations. In other words, the highest genetic gains from selection can be expected at sites 2 and 3 for D, H, and NN, and at site 1 for SUR (Table 3). However, our results may have been affected by the mortality rate at site 1, resulting from inter-tree competition, which may have affected growth traits such as D and H, decreasing the estimates of genetic parameters such as $h_m^2$, $h_w^2$, and $r_a$, because D is competition -sensitive (Leonardecz-Neto et al. 2003, Pavan et al. 2011). This could explain the inverse patterns of the estimated heritability and tree mortality rate. Under this condition, an efficient genetic selection at site 1 would be difficult, because D would be affected by a major environmental effect (competition), which would also affect the selection of genotypes for the best progeny for tree growth (Pavan et al. 2011). This aspect also has a negative effect on the selection of genotypes for growth control with regard to tree quality. For this purpose, the number of nodes (NN) is evaluated, which is related to the branching habit. The nodes are related with the wood quality for causing defects in the trunk, since most fractures in wood structures occur due to the presence of knots (Jansons et al. 2009, Corvalan-Vera 2020, Reyes-Esteves et al. 2022).

After selection, NN can be mitigated by silvicultural practices such as selective thinning and pruning. However, breeding allows a more effective long-term solution, since if sufficient genetic variation for NN is available (Vargas-Hernández et al. 2003), its expression can be controlled. Therefore, breeding programs include the control of traits associated with wood quality (Tong et al. 2013), as well as other traits related to branching, sinuosity and stem branching (Sotelo Montes et al. 2007a, Santos et al. 2010, Tung et al. 2010).

For D and H, the $h_m^2$ was higher than reported in a test with 36-month-old Guazuma crinita progenies, assessed at the same sites (Revilla-Chávez et al. 2022), where $h_m^2$ values were low for D (0.033-0.39, site mean = 0.369) and H (0.052-0.36, site mean = 0.34). This comparison indicated a higher potential for breeding by selection of the tested C. spruceanum than the G. crinita progenies.

Genetic ($r_g$) and phenotypic correlations ($r_p$) varied between pairwise traits at the sites (Table 4). Genetic correlations ($r_g$) were high and positive (≥ 0.6) between pairwise traits D, H, NN, and NB at each and all sites ($r_g$: 0.77-1.0), except between NN and NB at site 3 ($r_g$= 0.56). The phenotypic correlation ($r_p$) was generally high between pairwise traits at each and all sites ($r_p$: 0.6-0.89), except between D and NN at site 3 (0.59), D and NB at site 2 and the combined sites (0.48-0.57) and between H and NB at site 3 and combined sites (0.42-0.56), NN and NB at sites 2 and 3, and combined sites (0.36-0.55). Our findings corroborated results of Sotelo Montes et al. (2006) for the study species, who also reported high $r_g$ between D and H (0.89-0.91), and data of Sotelo Montes et al. (2003) and Tauchen et al. (2011), who found high $r_p\mathrm{}\mathrm{}$between D and H (0.85) and moderate $r_p$between NN and NB (0.45). The high $r_g$ between D and H indicated that the two traits can be considered as one in the selection process, since the selection for trees with higher D will also induce an increase in H, and since $r_g$ between both traits is largely influenced by the same pleiotropic genes (Vencovsky and Barriga 1992). This result is very favorable for indirect selection, since positive associations between traits indicate that selection on one trait can lead to indirect gains in another (Sotelo Montes et al. 2007a, Revilla-Chávez et al. 2022, Schoffen et al. 2023).

Table 4
Genotypic ($r_g$, below diagonal) and phenotypic ($r_p$, above diagonal) correlations for diameter (D), height (H), number of nodes (NN) and number of branches (NB) of each site and combined sites in tests with 38-month-old Calycophyllum spruceanum progenies

	D	H	NN	NB
Site 1
D		0.81	0.77	0.88
H	0.89		0.79	0.96
NN	0.68	0.80		0.97
NB	0.71	0.67	0.6
Site 2
D		0.88	0.78	0.74
H	0.87		0.87	0.89
NN	0.67	0.74		0.85
NB	0.57	0.65	0.55
Site 3
D		0.85	0.82	0.71
H	0.78		0.97	0.68
NN	0.59	0.68		0.56
NB	0.65	0.42	0.49
Combined
D		0.89	0.91	0.95
H	0.83		0.97	0.96
NN	0.65	0.64		1.0
NB	0.48	0.56	0.36

The progeny test was designed to serve two purposes: the ex situ conservation of Calycophyllum spruceamun combined with the production of seeds with some level of breeding. Thus, since the installed blocks were isolated from each other within the sites, we chose the strategy of selecting 75 of the 200 families (among) and one plant within each family (selection within) per block at each site (Table 5). Therefore, each block was transformed into a seed orchard, resulting in a total of 172-175 families per site and 199 families across the three sites for selection. This established an effective population size of 75 within the blocks, varying from 264 to 270 between sites, and totaling an effective population size of 472 for the three sites. This selection strategy achieved the genetic gains expected by selection among and within families, which varied between traits and sites. As expected, the possibility of genetic gains for the traits with highest heritability (D, H, NN and NB) was between 6.1 and 29.8%.

Table 5
Number of selected trees and families, group coancestry, and effective population size ($N_e$) of each site and combined sites in tests with 38-month-old Calycophyllum spruceanum progenies

	Site 1	Site 2	Site 3	Combined
Selected trees per block (1 tree/family)	75	75	75	75
Total number of selected trees per site	375	375	375	1125
Total number of selected families per site	172	173	175	199
Group coancestry per block: ${\mathrm{\Theta }}_b$	0.0067	0.0067	0.0067	0.0067
Group coancestry per site: ${\mathrm{\Theta }}_s$	0.0019	0.0019	0.0018	0.0011
Effective population size per block: $N_{e\left(b\right)}$	75	75	75	75
Effective population size per site: $N_{e\left(s\right)}$	265	264	270	472

Conclusions

Genetic variation among families was detected at each site and combined sites for the traits trunk diameter (D) and height (H) and number of branches (NB), as well as for number of nodes (NN) at sites 1 and 3 and combined sites, and for survival rate (SUR) at sites 1 and 2, which can be exploited by selection in breeding programs with 38-month-old C. spruceanum progenies.

The genotype - environment interaction for the traits D, H, NN, and SUR in 38-month-old trees is simple, so selection may be carried out at only one site.

The genetic control of the traits D, H, NN and NB, measured by heritability coefficients among and within families, varies from moderate to high, indicating the possibility of genetic gains by selection among and within families.

The genetic and phenotypic correlations between D, H, NN, and NB for each site and combined sites are positive and high, and selection for any of these traits will result in indirect genetic gains for the others.

The strategy of applying selection among and within families, within blocks and sites, serves two purposes: improvement by breeding and ex situ conservation, while the effective size of the remaining population can be exploited in other selection cycles.

Acknowledgments

The authors are indebted to the World Agroforestry Center (ICRAF) for allowing the use of field data underlying this article, to the Brazilian Federal Agency for Support and Evaluation of Graduate Education (CAPES) for granting JMRC a PhD scholarship and to the Brazilian Council for Scientific and Technological Development (CNPq), for awarding a Research Fellowship to AMS.

REFERENCES

• Boivin-Chabot S, Margolis H, Weber JC2004 Variation in coppice-shoot growth among provenances of Calycophyllum spruceanum Benth. in the Peruvian Amazon Basin. Forest Ecololgy and Management 198:249-260
• Canuto DSO, Silva AM, Freitas MLM, Sebbenn AM, Moraes MLT2016 Genetic variability in Myracrodruon urundeuva (Allemão) Engl. progeny tests. Open Journal of Forest 7:1-10
• Carmo ALM, Mazaratto EJ, Eckstein B, Santos AF2017 Associação de fungos com sementes de espécies florestais nativas. Summa Phytopatologica 43:246-247
• Corvalán-Vera P2020 Consideraciones silvícolas para la producción de postes en plantaciones de Pinus radiata D. Don en Chile. Revista Cubana de Ciencias Forestales 8:375-391
• Dávila-Lara A, Affinely M, Tribsch A, Díaz V, Comes HP2017 AFLP diversity and spatial structure of Calycophyllum candidissimum (Rubiaceae), a dominant tree species of Nicaragua’s critically endangered seasonally dry forest. Heredity 119:275-286
• Dawson IK, Lengkeek A, Weber JC, Jamandass R2009 Managing genetic variation in tropical trees: Linking knowledge with action in agroforestry ecosystems for improved conservation and enhanced livelihoods. Biodiversity and Conservation 18:969-986
• Hijmans RJ, Cameron SE, Parra JL, Jones PG, Jarvis A2005 Very high resolution interpolated climate surfaces for global land areas. International Journal of Climatology 25:1965-1978
• Holdridge LR1978 Ecologia basadas en zonas de vida. Instituto Interamericano de Ciencias Agrícolas, Turrialba, 235p
• Jansons A, Baumanis I, Haapanen M2009 Branch traits as selection criteria in Scots pine breeding in Latvia. Latvijas Lauksaimniecības Universitāte Rakski 23:45-56
• Leonardecz-Neto E, Vencovsky R, Sebbenn AM2003 Adjusting to plant competition in forestry tree progeny and provenance trials. Scientia Forestalis 63:136-149
• Lindgren D, Gea LD, Jefferson PA1996 Loss of genetic diversity monitored by status number. Silvae Genetica 45:52-59
• Menegatti RD, Mantovani A, Navroski MC2016 Genetic parameters for early growth traits in bracatinga progenies in Lages, SC, Brazil. Pesquisa Florestal Brasileira 36:235-243
• Orrego MDR, Bustamante GNR2017 Trabajabilidad de la madera de capirona (Calycophyllum spruceanum) procedente de plantaciones de la cuenca del río Aguaytia en la región de Ucayali-Perú. Revista for Perú 32:97-106
• Pavan BE, Paula RC, Perecin D, Candido LS, Scarpinati EA2011 Minimizing inter-genotypic competition effects to predict genetic values and selection in forestry genetic tests. Scientia Agricola 68:671-678
• Ramírez DU2016 Comportamiento fenológico preliminar de capirona en la provincia de San Martín. Instituto Nacional de Innovación Agraria, Lima, Perú, 2p
• Resende MDV2002 Genética biométrica e estatística no melhoramento de plantas perenes. EMBRAPA Informação Tecnológica, Brasília, 975p
• Resende MDV2016 Software Selegen - REML/BLUP: a useful tool for plant breeding. Crop Breeding and Applied Biotechnology 16:330-339
• Revilla-Chávez JM, Moraes MA, Pinchi-Ramirez MH, Sebbenn AM2022 Productivity, adaptability, and stability in Guazuma crinita progeny tests across three environments in the Aguaytia River Basin, Ucayali, Perú. Silvae Genetica 71:72-80
• Reyes-Esteves GI, López-Upton J, Velasco-García MV, Jiménez-Casas M2022 Genetic parameters of a progeny trial of Pinus greggii Engelmann ex Parlatore var. australis Donahu & López in the Mixteca Alta of Oaxaca, México. Revista Chapingo Serie Ciencias Forestales y del Ambiente 28:75-88
• Reynel C, Pennington RT, Pennington TD, Flores C, Daza A2003 Árboles útiles de la Amazonía peruana y sus usos. Tarea Gráfica Educativa, Lima, 509p
• Russel JR, Weber JC, Booth A, Powell W, Sotelo Montes C, Dawson IK1999 Genetic variation of riverine populations of Calycophyllum spruceanum in the Peruvian Amazon Basin, revealed by AFLP analysis. Molecular Ecology 8:199-204
• Saldaña CL, Cancan JD, Cruz W, Correa MY, Ramos M, Cuellar E, Arbizu CI2021 Genetic diversity and population structure of capirona (Calycophyllum spruceanum Benth.) from the Peruvian Amazon revealed by RAPD markers. Forests 12:1125
• Santos FW, Moraes MLT, Florsheim SMB, Lima IL, Silva JM, Freitas MLM, Sebbenn AM2010 Variação genética para caracteres anatômicos e retração volumétrica e sua correlação com a densidade básica da madeira em uma população base de Eucalyptus camaldulensis Dehn. Scientia Forestalis 38:159-170
• Schoffen VC, Cappa EP, Gauchat MW, Belaber EC2023 Genetic parameter estimation for Ilex paraguariensis St. Hill. in Argentina using spatial analysis. Crop Breeding and Applied Biotechnology 23:e46082344
• Sears R, Cronkleton P, Perez-Ojeda del Arco M, Robiglio V, Putzel L, Cornelius JP2014 Producción de madera en sistemas agroforestales de pequeños productores: Una justificación de política forestal a favor de los pobres en el Perú. Center for International Forestry Research (CIFOR), Bogor, 8p
• Sebbenn AM, Freitas MLM, Zanatto ACS, Moraes E2009b Seleção dentro de progênies de polinização aberta de Cariniana legalis Mart. O. Ktze (Lecythidaceae), visando à produção de sementes para recuperação ambiental. Revista do Instituto Florestal 21:27-37
• Sebbenn AM, Freitas MLM, Zanatto ACS, Moraes E, Moraes MA2009a Comportamento da variação genética entre e dentro de procedências e progênies de Gallesia integrifolia Vell. Moq. para caracteres quantitativos. Revista do Instituto Florestal 21:151-163
• Sebbenn AM, Zanatto A, Freitas ML, Sato A, Ettori L2005 Genetic variation in Araucaria cunninghamii provenances in Luiz Antonio-SP, Brazil. Crop Breeding and Applied Biotechnology 5:435-442
• Sotelo Montes C, Beaulieu J, Hernandez RE2007a Genetic variation in wood shrinkage and its correlations with tree growth and wood density of Calycophyllum spruceanum at an early age in the Peruvian Amazon. Canadian Journal of Forest Research 37:966-976
• Sotelo Montes C, Beaulieu J, Hernandez RE2007c Genetic variation in wood mechanical properties of Calycophyllum spruceanum at an early age in the Peruvian Amazon. Wood and Fiber Science 39:578-590
• Sotelo Montes C, Hernandez RE, Beaulieu J, Weber JC2006 Genetic variation and correlations between growth and wood density of Calycophyllum spruceanum at an early age in the Peruvian Amazon. Silvae Genetica 55:217-228
• Sotelo Montes C, Hernandez RR, Beaulieu J, Weber J2008 Genetic variation in wood color and its correlations with tree growth and wood density of Calycophyllum spruceanum at an early age in the Peruvian Amazon. New Forest 35:57-73
• Sotelo Montes C, Vidaurre H, Weber J2003 Variation in stem-growth and branch-wood traits among provenances of Calycophyllum spruceanum Benth. from the Peruvian Amazon. New Forest 26:1-16
• Sotelo-Montes C, Hernandez RE, Beaulieu J2007b Radial variation in wood density and correlations with growth of Calycophyllum spruceanum at an early age in the Peruvian Amazon. Wood and Fiber Science 39:377-387
• Tauchen J, Lojka B, Hlasna-Cepkova P, Svobodova E, Dvorakova Z, Rollo A2011 Morphological and genetic diversity of Calycophyllum spruceanum (Benth) K. Schum (Rubiaceae) in Peruvian amazon. Agricultura Tropica et Subtropica 44:212-218
• Taylor L2005 The healing power of rainforest herbs: A guide to understanding and using herbal medicinals. Square One Publishers, New York, 359p
• Tong Q, Duchesne I, Belley D, Beaudoin M, Swift E2013 Characterization of knots in plantation white spruce. Wood and Fiber Science 45:84-97
• Tung EDC, Freitas MLM, Florsheim SMB, Lima IL, Longui EL, Santos FW, Moraes MLT, Sebbenn AM2010 Variação genética para caracteres silviculturais e anatómicos da madeira em progênies de Myracrodruon urundeuva (Engler) Fr. Allem. Scientia Forestalis 38:499-508
• Ugarte-Guerra LJ, Domínguez-Torrejón G2010 Índice de Sitio (IS) de Calycophyllum spruceanum Benth. en relación con la altura dominante del rodal en ensayos de plantación en la Cuenca del Aguaytía, Ucayali, Perú. Ecología Aplicada 9:101-111
• Vargas-Hernández JJ, Adams WT, Joyce DG2003 Quantitative genetic structure of stem form and branching traits in Douglas-fir seedlings and implications for early selection. Silvae Genetica 52:36-44
• Vencovsky R, Barriga P1992 Genética biométrica no fitomelhoramento. Revista Brasileira de Genética, Ribeirão Preto, 486p
• Weber JC, Sotelo Montes C2005 Variation and correlations among stem growth and wood traits of Calycophyllum spruceanum Benth. from the Peruvian Amazon. Silvae Genetica 54:31-41
• Weber JC, Sotelo Montes C, Ugarte J, Simons T2009 Phenotypic selection of Calycophyllum spruceanum on farms in the Peruvian Amazon: Evaluating a low-intensity selection strategy. Silvae Genetica 58:272-279

Publication Dates

• Publication in this collection
  08 Mar 2024
• Date of issue
  2024

History

• Received
  18 Oct 2023
• Accepted
  23 Dec 2023
• Published
  28 Dec 2023