Evidence of low-levels of natural introgression between Pinus jaliscana and Pinus oocarpa in an open-pollinated progeny trial using near-infrared spectroscopy

There have been no reports of natural introgression between Pinus jaliscana and Pinus oocarpa where they occur sympatrically in the highlands of western Mexico. However, a small number of putative hybrid trees were observed in a progeny trial of 57 open-pollinated families of P. jaliscana at De Rust, South Africa that represented all nine known natural populations of the species in its native geographic range. The objectives of the study were to confirm hybridity using traditional botanical analysis of needle traits (no cones were available) and once admixtures were confirmed to develop near infrared spectroscopy (NIRS) models to define the spectral signature of pure P. jaliscana and the hybrid. The calibration models were then run on a large subsample of 260 trees that served as a validation data set to predict the range-wide occurrence of introgression in P. jaliscana. The calibration data set for the NIRS included 57 pure P. jaliscana and 31 hybrid trees identified in the trial that were morphologically separated using canonical analysis based on four external needle traits: number of needles per fascicle, needle length, needle width, fascicle sheath length and three needle color traits: hue, value and chroma. NIRS models were developed using 1 g of ground, oven-dried needles/tree that were scanned at visible and NIR wave lengths between 400 and 2500 nm. The raw spectral data, log(1/reflectance) were processed using standard normal variate and de-trending transformations. The NIRS calibration model identified six hybrid trees in the validation population for an overall species admixture rate of 2.3% versus 7.7% based on subjective scoring in the field. The implications of P. oocarpa introgression in P. jaliscana in natural stands in western Mexico are discussed. NIRS provides a quick way to obtain reasonable estimates of species admixtures in progeny trials and is an inexpensive surrogate to using molecular markers. NIRS seems most practical when species are very closely related and/or there is unbalanced representation of parental species in progeny tests to make comparisons with putative hybrids difficult. (C) 2013 Elsevier B.V. All rights reserved.