Agricultural Water Management 121 (2013) 102–112 Contents lists available at SciVerse ScienceDirect Agricultural Water Management j ourna l ho me p ag e: www.elsevier.com/locate/agwat Response of taro (Colocasia esculenta L. Schott) landraces to varying water regimes under a rainshelter T. Mabhaudhi a,∗ , A.T. Modi a , Y.G. Beletse b a Crop Science, School of Agricultural, Earth & Environmental Sciences, University of KwaZulu-Natal, Pietermaritzburg, South Africa b Agricultural Research Council, Roodeplaat, Vegetable and Ornamental Plant Institute (VOPI), Private Bag X293, Pretoria 0001, South Africa a r t i c l e i n f o Article history: Received 3 August 2012 Accepted 12 January 2013 Available online 28 February 2013 Keywords: Drought Neglected underutilised species Water-use efficiency Yield a b s t r a c t Taro [Colocasia esculenta (L.) Schott] is an underutilised crop in sub-Saharan Africa due to lack of agro- nomic research on it. There is no information describing water-use and drought tolerance of local taro landraces. Therefore, the objective of this study was to evaluate growth, yield and water-use of three South African landraces of taro under varying water regimes. Three taro landraces [Dumbe Lomfula (DL), KwaNgwanase (KW) and Umbumbulu (UM)] were planted in a rainshelter (14, October, 2010 and 8, September, 2011) at Roodeplaat, Pretoria, South Africa. Three levels of irrigation [30%, 60% and 100% crop water requirement (ETa)] were applied three times a week using drip irrigation. Emergence, plant height, leaf number, leaf area index (LAI) and stomatal conductance were measured in situ. Root length, fresh and dry mass were obtained by destructive sampling. Yield, yield components and water-use efficiency were determined at harvest. Taro landraces showed slow and uneven emergence. Stomatal conductance was respectively, 4% and 23% lower at 60% and 30% ETa relative to 100% ETa. Such a decline was clearer in the UM landrace, suggesting greater stomatal regulation in the UM landrace compared with KW and DL landraces. Plant growth parameters (plant height, leaf number and LAI) were shown to decrease by between 5% and 19% at 60% and 30% ETa, respectively, evapotranspiration relative to 100% ETa. The KW and DL landraces were shown to decrease the most while the UM landrace had moderate reductions in growth. Taro yield was 15% and 46% higher at optimum irrigation relative to 60% ETa and 30% ETa, respec- tively. Water-use efficiency was relatively unchanged (0.22–0.24 kg m -3 ) across varying water regimes. On average, the UM landrace had 113% higher WUE than the KW landrace. These findings can be used to differentiate the landraces on the basis of potential drought tolerance. © 2013 Elsevier B.V. All rights reserved. 1. Introduction Taro (Colocasia esculenta L. Schott) is a major root crop of the Araceae family with wide distribution in the tropics and subtropics (Lebot, 2009). It is among the oldest crops known to man with a history dating back to more than 10,000 years (Rao et al., 2010). However, the crop remains underutilised in much of the world, including South Africa, due to lack of information. The widely held perception that taro is one of the least water efficient crops (Uyeda et al., 2011) may, in part, explain its current low levels of utilisation. It therefore comes as no surprise that information describing water- use of taro and possible drought tolerance is scarce. Only recently has a major project been commissioned where one of the objectives is to breed for drought tolerance in taro (International Network of Edible Aroids (INEA) http://www.ediblearoids.org). ∗ Corresponding author. Tel.: +27 33 260 5447; fax: +27 33 260 5073. E-mail addresses: tmabhaudhi@gmail.com, tafadzwanashemabhaudhii@yahoo.com (T. Mabhaudhi), modiat@ukzn.ac.za (A.T. Modi), beletsey@arc.agric.za (Y.G. Beletse). There were few reports in the literature describing drought tol- erance of taro and its water-use (Sivan, 1995; Sahoo et al., 2006; Uyeda et al., 2011). Sivan (1995) studied drought tolerance in two dasheen and eddoe taro varieties, as well as tannia (Xanthosoma sagittifolium) and observed that stomatal conductance, leaf num- ber and leaf area of both cultivars all decreased in response to water stress. In a separate study, Sahoo et al. (2006) subjected a taro hybrid to water stress using polyethylene glycol (PEG) and observed significant differences in plant growth parameters of height, leaf number and area as well as minimum yield reduction in response to water stress. Elsewhere, Uyeda et al. (2011) evaluated the response of three commercial taro varieties to five irrigation rates based on reference evapotranspiration (ET o ) and found that irrigating taro at 150% of ET o could maximise yield. Sahoo et al. (2006) went on to conclude that development of drought toler- ant taro cultivars was possible while Uyeda et al. (2011) stated that upland taro varieties may be adapted to water-limited produc- tion. Therefore, evaluating responses of previously unstudied taro landraces to water stress may aid in identifying genotypes with drought tolerance and suitability for production in water-limited areas. 0378-3774/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.agwat.2013.01.009