GEOMETRIC RELATIONSHIPS BETWEEN WHITEFLY FEEDING-BEHAVIOR AND VASCULAR BUNDLE ARRANGEMENTS

This study revealed strong evidence that nymphs of the silverleaf whit efly, Bemisia argentifolii Bellows and Perring, are obligate feeders o n vascular bundles and that there are large differences between differ ent host plants as to the availability of vascular bundles to silverle af whitefly nymphs. The relationship between nymphs and leaf vascular bundles was studied using 1) leaf sectioning and 2) techniques of leaf clearing of intact leaves. A geometric model is presented of the feed ing relationship of vascular bundle-using homopterans. The relative ab undance of vascular bundles was examined in six species of host plants that varied from highly preferred to tolerably acceptable. Included i n order of acceptance were cantaloupe, cotton, hibiscus, broccoli, lan tana and lettuce. The length of vascular bundle per 1.0 mm(2) of leaf surface ranged from about 10 mm in cantaloupe to 2.8 mm in lettuce. Sa livary sheaths were found to connect with vascular bundles in 100% of the intact nymphs examined by the staining and clearing technique. How ever only 64% of those examined by the sectioning technique appeared t o be connected to vascular bundles. This indicates that the sectioning technique leads to a high rate of error, causing an underestimation o f the importance of direct contact with vascular bundles. About 50% of epidermal stylet penetrations were through epidermal cells; the remai ning 50% went through intercellular junctions. On cotton leaves, the d istance between the point of labial contact with the leaf surface and the nearest point of the vascular bundle rarely exceeded 60 mu m. Our studies show that while 50% of lettuce leaf-surface was beyond 60 mu m of a vascular bundle, only 10% of cantaloupe leaf surface area was ou tside of the 60 mu m range. In cotton, mean distance from labium to th e nearest point of the vascular bundle was 40.9 mu m (SEM=2.66, N=50, range 0-80 mu m). Over 98% of all salivary sheaths went to minor veins (78% to single-filament vascular bundles, nearly 20% to double filame nt bundles). Fewer than 2% went to bundles with 3 or more filaments.