A new potential method to estimate abundance of small game species Cletah Shoko 1 *, Mhosisi Masocha 1 and Timothy Dube 2 1 Department of Geography and Environmental Science, University of Zimbabwe, PO Box MP 167, Harare, Zimbabwe and 2 Discipline of Geography, School of Agricultural, Earth and Environmental Sciences, University of KwaZulu-Natal, P/Bag X01, Scottsville, Pietermaritzburg, 3209, South Africa Abstract This study tested the hypothesis that nonlinear transects, such as the L-, V- and W-shaped transect, which are widely used in soil sampling but rarely in ecology, may perform better in capturing habitat heterogeneity poten- tially resulting in more accurate small game species estimates, than straight-line transects. To test this, we computed and compared the Normalized Difference Vege- tation Index (NDVI) captured by the straight-line, L-shaped, V-shaped and W-shaped transects for Kyle Game Reserve (Zimbabwe), inside the home ranges of two small game species, namely species A and species B. These transects differed in geometry but had the same length. We also performed simulations in a geographic information system environment to compare the detection ability of the different transect geometry. One-way analysis of variance followed by Tukey’s honest significant difference test was used to test for significant differences in the mean number of detections and NDVI variance captured by each transect geometry. Simulations results indicate that for both home ranges, the V-shaped transect resulted in significantly higher detections than the widely used straight-line transect. Nonlinear transects also yielded higher NDVI variances than the straight-line transect in both home ranges. Key words: ASTER, experimental design, Kyle Game Reserve, Normalized Difference Vegetation Index, transect geometry R esum e Cette  etude a test e l’hypoth ese selon laquelle des transects non lin eaires (en forme de L, de V ou de W), qui sont fr equemment utilis es pour des  echantillonnages de sol mais rarement en  ecologie, pourraient donner de meilleurs r esultats pour capter l’h et erog en eit e de l’habitat et donc aussi des estimations plus exactes de petites esp eces de gibier que des transects en ligne droite. Pour le v erifier, nous avons calcul e et compar e l’Indice de v eg etation par diff erence normalis ee(NDVI) obtenu pour des transects en ligne droite, en L, en V et en W, dans la R eserve de Faune de Kyle (Zimbabwe), dans l’aire de r epartition de deux petites esp eces de faune, reprises comme esp ece A et esp ece B. La g eom etrie de ces transects  etait diff erente mais ils avaient la m^ eme longueur. Nous avons aussi r ealis e des simulations dans un environnement de Syst eme d’information g eo- graphique (SIG) pour comparer la capacit e de d etection des diff erentes g eom etries de transects. Une analyse de la variance  a un seul facteur, suivie d’un test HSD de Tukey, a servi  a tester les diff erences significatives du nombre moyen de d etections et la variance du NDVI capt ee par chaque g eom etrie de transect. Les r esultats des simulations indiquent que, pour les deux aires de r epartition, le transect en V donnait des d etections significativement plus nombre- uses que le transect en ligne couramment utilis e. Les transects non lin eaires r ecoltaient aussi des variantes de NDVI plus  elev ees que le transect en ligne droite dans les deux aires de r epartition. Introduction Small game species are a vital component of ecosystems. They play an important role in the transfer of energy and matter as well as serving as secondary consumers and prey for carnivores (Schmidt et al., 2005). Thus, they can exert a significant influence on predator population cycles (Sieg, 1987) and as important bio-indicator of ecosystem health (Fricke, Kempema & Powell, 2009). Therefore, their *Correspondence: E-mail: kiletashoko@gmail.com © 2015 John Wiley & Sons Ltd, Afr. J. Ecol. 1