The study area is an experimental plot established by a team of research scientists from the Forestry Research Institute of Ghana in Tano-Nimiri, a degraded forest reserve, which is undergoing reforestation through col- laboration between Samartex Timber and Plywood Company limited and surrounding communities. The area is a concession for Samartex located in Samreboi, a town in the western region of Ghana which is within the wet evergreen forest type in Ghana. It lies between altitudes of between 60 m and 180 m with peaks rising to as high as 260 m (Figure 1). The annual rainfall is between 1750 mm and 2000 mm (Hall & Swaine, 1981; Asankragwa, 2006). The average low temperature of the area is 21˚C and average high temperature of the area is 32˚C. Rela- tive humidity is also between 70% and 85% perseason (Forest Management Unit 10, 2001).

The experimental site is a ten (10) hectare area, laid out in a randomized complete block design, with four blocks. Each block was divided into 10 plots with varying percentages of Khaya grandifoliola C.DC (Kg), Khaya ivorensis A. Chew (Ki), Heritieria utilis Sprague (Hu), Terminalia superba Engl and Diels (Ts) and En- tandrophragma angolense Welw. (Ea). All the species grow naturally in the area except K. grandifoliola. Each plot had a dimension of about nineteen meters by ninety meters (19 m × 90 m) with a spacing of three meters (3 m) between trees. In each plot, a total of 220 individual trees were planted. The present data were collected when the trees were two years old.

The key species for this study were Khaya grandifoliola and Khaya ivorensis, which were planted at relative densities of 100% s (pure stand), 60%, 50%, 40%, 20% and 10% per plot with an even mixture of the 3 compan- ion species (Heritieria utilis Sprague, Terminalia superba and Entandrophragma angolense). A systematic ran- dom sampling method was used to collect the data. In each block, an average of thirty (30) trees per plot were selected and assessed for each Khaya species. On the other hand 15 trees were assessed for each of the compan- ion trees. The growth measurements taken were total tree height (Ht), diameter at breast height (DBH; measured at a standard height of 1.3 m) and height at first fork (HtF) for the Khaya species but only diameter and height were measured for the companion species.

Damage by H. robusta was assessed by recording total shoots attacked (TSA), total number of shoots with fresh attack (TFA), total number of dead shoots (TDS), number of total shoots (TS) sprouted in response to Hypsipyla attack, and the length of the longest dead shoot (LDS). In all about 746 trees were sampled for Khaya grandifoliola and 467 for Khaya ivorensis.

An analysis of variance (ANOVA) using the General Linear Model (GLM) of the SPSS statistical package (ver- sion 16) was used to test for differences between the means of the parameters assessed at the five percent signi- ficance level (P < 0.05). The model chosen allows for pair-wise and multiple comparisons of the treatments. Tree volume measurements were calculated using Newbould, 1967: V = 1/2 (B_dbh × Ht), where V is volume, B_dbh is the basal area of diameter at breast height and Ht is tree height.

Khaya grandifoliola demonstrated a better growth performance than Khaya ivorensis in all planting densities with 10% showing the best growth performance for K. grandifoliola (6.44 cm diameter, 4.90 m height and 0.0105 m³ volume) and 60% for K. ivorensis (5.27 cm diameter, 4.38 m height and 0.0057 m³ volume) (Table 1). No differences were observed in growth with relative Khaya density for both diameter and height of K. grandifoliola (P = 0.172 and 0.463 respectively, Table 2), but for K. ivorensis, growth differences with relative Khaya density were observed for both diameter and height (P = 0.000, Table 3).

Generally, Terminalia superba had a greater growth rate than the other mixed species, with an average di- ameter of 7.2 cm and average height of 4.7 m. Entandrophragma angolense had an average diameter of 3.1 cm and average height of 3.1 m, while, Heritieria utilis had an average diameter of 2.0 cm and average height of 2.3 m. Khaya grandifoliola had an average diameter of 6.2 cm and height of 4.8 m and Khaya ivorensis had an av- erage diameter of 4.5 cm and height of 3.9 m (Figure 2 and Figure 3).

In general K. ivorensis showed less forking than K. grandifoliola, except at 60% density (Figure 4) in the mixed plantings two years after planting in the field. There were significant differences among densities for both K. grandifoliola and K. ivorensis in mean height at first fork (P = 0.010 and P = 0.018, respectively, Table 2 and Table 3). The density which had the best height to first fork was 100% planting density at 2.16 m for K. grandifoliola and 60% planting density at 1.58 m for K. ivorensis (Figure 4).

The number of shoots with fresh attack recorded was very low for both species, with K. grandifoliola recording a range of 0.017 (40% density) to 0.098 (at 100% density) and a range of 0.00 (at 60% density) to 0.053 (at 100%

Note: Standard errors for means given in brackets.

density) for K. ivorensis (Table 4). Statistically, there were no significant differences among densities in fresh shoot attacks for either Khaya species (Table 2 and Table 3).

The results two years after planting in the field revealed that K. grandifoliola responded to Hypsipyla robusta attack with more branches than K. ivorensis (Figure 5). The number of shoots of K. grandifoliola and K. ivoren- sis attacked by Hypsipyla robusta was higher in K. grandifoliola than in K. ivorensis. The highest number of shoots attacked for K. grandifoliola was 1.39 obtained at 50% density for the mixed-species stands and lowest number of shoots attacked was 0.89 (at 10% density) (Figure 6). The highest number of shoots attacked for K. ivorensis was 0.89 (at 100% density) and lowest was 0.32 (at 20% density). For K. ivorensis, the number of total shoots ranged from 1.43 (at 20% planting density) to 2.16 (at 60% planting density) while for K. grandifoliola the number of total shoots ranged from 2.47 (at 10% density) to 3.38 (at 100% planting density).

K. grandifoliola recorded the lowest number of dead shoots at 0.15 (at 40% planting density) and the highest at 0.74 (at 60% planting density) and the length of the longest dead shoot at 10% planting density (Table 5). The number of dead shoots for K. ivorensis ranged from 0.085 (at 20% planting density) to 0.37 (at 100% planting density) (Table 5).

Note: Standard errors for means given in brackets.

Note: Standard errors for means given in brackets.

Two years after planting Khaya grandifoliola and Khaya ivorensis in a mixed planting of different densities with Terminalia superba, Entandrophragma angolense and Heritieria utilis which are all important commercial tim- ber trees in the tropics, differences were displayed for growth in diameter between the planting densities with the 10% density having the best increase in girth for K. grandifoliola and 60% for K. ivorensis. K. grandifoliola, however, did not demonstrate differences between mixed-planting densities for growth in height. This trend is similar to what is reported for monocultures and mixed plantations of Eucalyptus and Acacia (Khanna, 1997; Parrotta, 1999), Monocultures and mixed plantations of poplar species and Alus glotinosa (Hansen & Dawson, 1982). Also study conducted in the dry semideciduous forest in Ghana for mixed plantation of K grandifoliola with Terminlia superba and Cedrela odorata followed a similar trend (Opuni-Frimpong et al., 2013). The rela- tive slow growth rate of the companion species might have given advantage to the faster growing K. grandifoli- ola, exposing it to more light and could photosynthesize more effectively than the other species, leading to quicker growth rate (Hawthorne, 1995; Poorter et al., 2004; Petit & Montagnini, 2006). On the other hand, there were significant differences between planting densities for growth in height and diameter of K. ivorensis trees, with the 60% K. ivorensis mixed stands having the best growth. This reveals that different species may require different treatments in mixed density planting. Even though the two species are of the same genus, they per- formed better under different mixed densities and thus each species requires some compatible level of mixtures to enhance the efficient use of resources in mixed-stands for productivity as observed by Forrester et al. (2006) and Kelty (2006). Height at first fork was recorded for both Khaya species to determine the harvestable bole length for the trees and again in this case each species performed better differently in density mixed planting.

The study conducted in the wet evergreen forest zone of Ghana, which is the natural range of K. ivorensis (Hall & Swaine, 1981; Oteng-Amoako, 2006), observed that K. grandifoliola, which naturally occurs in dry semideciduous forests (Poorter et al., 2004), showed a better growth performance than K. ivorensis in all of the six mixed-densities planted. This could have been as a result of K. grandifoliola’s ability to adapt to more fa- vourable environmental conditions in an ecological zone which has a better rainfall regime leading to its faster growth. Opuni-Frimpong et al. (2008a) also suggested that improved growth characteristics exhibited by Khaya anthotheca (genetically similar to K. grandifoliola) in a relatively moist forest as compared to its natural range of drier forests could have been as a result of better environmental conditions provided. This proves that species may have different use-efficiency for relevant resource like good rainfall pattern which could affect plants growth and species distribution (Grijpma, 1976; Swaine, 1996; Engelbrecht et al., 2007).

Silvicultural interventions to manage mahogany shoots borer (Hypsipyla robusta) problems and maximize growth of trees for timber in mahogany plantations, try to interfere with the mahogany shoot borer’s ability to locate the host plant, reduce host suitability, encourage natural enemies and assist recovery of the trees after at- tack (Hauxwell et al., 2001). These measures include planting vigorous seedlings at good sites together with other plant species that may physically obstruct or may release chemicals that interfere with the chemical cues that help the shoot borer to locate the host plant (Opuni-Frimpong et al., 2005; Hauxwell et al., 2001; Griffiths, 2001).

Branching is one of the main features that expresses the effects of H. robusta attack. This occurrence has also been observed by Griffiths (2001); Nair (2001); Opuni-Frimpong et al. (2008b), whose work ascertained that frequent attacks on young plants generally lead to poor quality timber. The results of this study also showed that branching occurred in the mahogany trees in all six planting densities (pure and mixed-species stands) resulting in multiple shoots. The number of new shoots produced showed the mahogany’s ability to recover from shoot borer attack. K. grandifoliola had more shoots than K. ivorensis which implied that K. grandifoliola had more H. robusta attack. For pure stands (100% planting density of each species) K. grandifoliola had 41% more shoots than K. ivorensis. This could possibly be because K. grandifoliola is more vulnerable to the mahogany shoot borer’s attack than K. ivorensis. Similar findings were made by Opuni-Frimpong et al. (2008a), whose study on African mahogany species showed that K. anthotheca was more susceptible to H. robusta attack than K. ivoren- sis. Taller mahogany trees has been observed to attract higher Hypsipyla attacks in plantations (Cipiao et al., 2009; Perez-Salicrup & Esquivel, 2008; Opuni-Frimpong et al., 2008, 2013) and thus K. grandifoliola with its fast growth in the mixed plantation was more susceptible to Hypsipyla attack.

The lowest levels of H. robusta attack were observed in the 20% and lower planting densities for both Khaya species. This observation corroborates earlier reports that lower densities of mahogany in mixture stands tend to make it difficult for the mahogany shoot borer to locate the Khaya species, consequently reducing the host con- centration of the pest (Opuni-Frimpong, et al., 2005; Hauxwell et al., 2001; Griffiths, 2001; Kelty, et al., 2006).

Even though the effect of shade was not explicitly examined in this study, it was observed that low levels of attack by H. robusta were recorded in areas where K. grandifoliola and K. ivorensis were predominantly shaded by other plants in the mixture stands; this observation corresponds to a study conducted by Opuni-Frimpong et al. (2008b), which examined the effect of canopy shade on some mahogany species. The results of that study in- dicated that, canopy shade decreased H. robusta attack levels in the mahogany species studied; however, it also led to a reduction in growth of the species concerned as observed in this study for K. ivorensis. The number of dead shoots (die-back) for both Khaya species gives us an indication of the levels and extent of H. robusta at- tacks on the mahogany trees. It is also an indication of the mahogany’s ability to recover from the incidence of shoot borer attack and produce new shoots. According to Hauxwell et al. (2001), Newton et al. (1999), and Opuni-Frimpong (2006), this observation may be as a result of the Khaya species mechanism of self-pruning which is attributed to the species.

Of the six densities studied, mixture stands with equal percentages of all species (20% density) and the 10% density had the lowest level of H. robusta attacks. Mixed species plantation according to Opuni-Frimpong et al. (2013), Kelty (2006), Haux-well et al. (2001), Watt (1994), and Mayhew and Newton (1998), may help in re- ducing the incidence of H. robusta infestation in mahogany plantations. This expression of low levels of H. ro- busta infestation at lower densities of Khaya species may be attributed to stronger interference in chemical cues from companion species disrupting Hypsipyla ability to easily identifying host trees (Matsumoto & Kotulai, 2000). Again the companion species may have made available more diverse habitats to support greater popula- tion of natural enemies of Hypsipyla and potentially by providing a refuge for natural enemies leading to low at- tacks (Watt, 1992; Parker et al., 2013).