Insect samples were collected at Kelso dunes over a nine-day period between April 24th and May 2nd, 2013. Insects were sampled daily using pan traps (approximately 18 cm in diameter) set at ground-level along an east-west axis parallel to Kelso Dunes. Thirty pairs of pan traps were set along two parallel 45 m transects (transects were 10 m apart) with alternating blue, yellow, and white traps approximately every 3 m using the NSERC-CANPOLIN protocol (http://www.uoguelph.ca/canpolin). Pan traps were paired so that each replicate had one pan trap under the southern portion of a L. tridentata canopy, halfway between the base of the shrub and the drip-line, and within a patch of annual plants. The other pan traps were deployed 2 m south of each paired shrub in an adjacent open microsite, also with annual plants present (see Appendix A; Fig. A2). Open microsites were located two metres from the drip-line of shrubs because this was on average the maximum distance possible without being within a two metre radius of another shrub (Ruttan pers. obs). Pan traps were half-filled with a solution of soapy water prepared by mixing five drops of unscented dish detergent per litre of water (for protocol, see: http://www.uoguelph.ca/canpolin). Pan traps were set out by 9:00 a.m. and collected at 5:00 p.m. daily targeting typical peak insect activity (http://www.uoguelph.ca/canpolin). All samples were collected on sunny days with no precipitation. Samples were collected from each pan trap replicate and stored in vials of 70% ethanol. Insects were then sorted from samples and identified to the family level for ease of identification using Goulet and Huber (1993) and Borror et al. (1989). Following identification, insects were categorized into their primary functional groups, including pollinators (mostly bees), herbivores, granivores, parasites, nec- tarivores (that contribute only marginally to pollination), and others.