Spatial distribution of roots of pearl millet on sandy soils of Niger

Root sampling in crop stands of low planting density requires reliable information on horizontal distribution of roots. This applies particularly to pearl millet in the Sahel, which is sown at a rate of less than two pockets of seed per m2. The objective of this study was to investigate the spatial variability of root length density (RLD) among sampling positions in an improved management system with ridging and under traditional sowing. RLD between ridges (bR) was lower compared to sampling positions within ridges (wR) at soil depth layers from 0 to 80 cm soil depth. We found a highly significant, positive correlation between the sum of the root length (RL) of four sampling dates (tillering, booting, flowering, and maturity) with shoot dry mass (SDM) at maturity. The square of the correlation coefficient was highest when calculation of RL was based on RLD at all four sampling positions. While SDM exhibited significant differences among three pearl millet varieties, sole root sampling wR at a lateral distance of 60 cm relative to the pocket would not allow for the detection of varietal differences in RL, while all other sampling positions did. The correlation between RL and SDM was considerably improved if information of RLD bR was included. Under traditional sowing, RLD directly under the plant was lower compared to sampling positions at lateral distance 25 and 50 cm from the centre of the pocket, but this effect of sampling position was not significant. RLD estimates within deeper soil layers were not systematically affected by direction and lateral distance. To obtain accurate information about depth of rooting and RL under traditional sowing, samples should be taken from lateral distances between 20 and 40 cm from the pocket centre.