A total of 30 Caesalpinia echinata (pernambuco) sticks were ranked based on their suitability for making high quality bows and were assigned to one of the three following categories: 0=very poor to poor, 1=good to very good, and 2=excellent. From the end of each stick a sample was cut for wood property and near infrared (NIR) spectroscopic analysis. Wood properties measured included air-dry density, extractives content, microfibril angle, stiffness and wood color. NIR spectra were evaluated by principal component analysis (PCA) and on the PC scores. Poor quality samples were discriminated from those of good to very good and excellent quality; however, samples from the two higher quality groups were mixed. Based on relationships observed between PC scores and wood properties, we suggest that, of the measured properties, density and stiffness were the most important in sample discrimination based on quality. Samples ranked in the excellent category had high average density (1119 kg m-3) and stiffness (25.2 GPa) and relatively low extractives content (21.2%) compared to samples in the very poor to poor category (density= 938 kg m-3, stiffness=18.9 GPa and extractives content=24.9%).

Caesalpinia echinata (pernambuco or pau-brasil), is recognized as the premier wood for manufacturing stringed instrument bows. Owing to limited supplies and concerns regarding species survival, interest exists in establishing pernambuco plantations to provide future bow-quality wood. For native forest-and plantationgrown woods we examined several wood properties considered important in determining bow quality including basic density, modulus of elasticity (MOE), modulus of rupture (MOR) measured using static bending samples, air-dry density, microfibril angle (MFA), and stiffness using SilviScan. Color, extractives content and loss tangent (tan δ) were measured for a subsample of the static bending samples. Finally, the samples were also ranked based on their potential for manufacturing high-quality bows (0 = poor, 1 = good and 2 = excellent) by an experienced bow maker. No evidence of differences between means for density, MOE and MOR for native, and 25-and 30-year-old plantation-grown pernambuco was observed; however, when sorted based on quality, the excellent group had higher density, MOE and MOR. MFA and tan δ were low, especially for native forest samples. Extractive contents were low for plantation samples; 5.7% and 12.7% respectively for the 25-and 30-year-old samples, compared to the native forest samples (set 1 = 23%, set 2 = 22.5 %) and few samples had heartwood. Overall, plantation-grown samples provided promising results in terms of their quality.

Near infrared (NIR) spectroscopy, coupled with multivariate data analysis, is proposed as a rapid and effective analytical method for evaluating the quality of pernambuco (Caesalpina echinata Lam.) wood for making bows for stringed instruments. For this purpose, a set of 30 pernambuco sticks were ranked based on their suitability for making high-quality bows and they were assigned to one of the following categories: 0 = very poor to poor, 1 = good to very good and 2 = excellent. Considering the low number of samples in the poor category, the classification study focused on the discrimination between samples of the two higher quality groups. Linear discriminant analysis (LDA) was applied to the NIR data as a classification technique and in order to ensure a more appropriate ratio between the number of objects (samples) and number of variables (absorbance at different wavenumbers), LDA was preceded by feature selection. Based on LDA, 100% of the samples were correctly classified and 92.6% of the samples were correctly predicted by the cross-validation procedure.

Caesalpinia echinata Lam. (pernambuco or pau-brasil) is recognized as the premier raw material for manufacturing stringed instrument bows. Several studies have identified properties considered important in determining the suitability of pernambuco wood for bow manufacture including density, modulus of elasticity (MOE), and, possibly, microfibril angle (MFA). No research has been conducted on how these properties vary within individual trees; however, an understanding of how pernambuco wood properties vary within trees is important as it may assist in the identification of trees or provenances most suited for the establishment of plantations, aid in developing an understanding of management practices on wood property variation for plantation-grown pernambuco and also facilitate the identification of regions within trees that possess optimal properties for bow manufacture. Radial variation in density, MFA and MOE was examined using SilviScan for three radial strips representing differing levels of wood quality in terms of perceived suitability for making high-quality bows. The lowest quality sample showed considerable radial variation compared to the higher quality samples for all properties and it also had the lowest average density. It was not possible to identify a strong pith to bark trend for any of the wood properties examined.