The graph just below illustrates the historical trend toward 650c (26") wheels from 700c in the Kona race. Why did 650c grow to a rate of relative equivalence with its larger cousin through the mid 1990s? Because when Quintana Roo introduced its Superform in 1989—the first true tri-bike which incorporated, among other things, steeper seat angles—the bike came with 650c wheels. Copies of that bike which had 700c wheels—Specialized made one—were a commercial failure. Copies which did have 650c wheels did well. Cannondale and Kestrel were two of the more successful examples.
It was therefore not unlike decisions of today to build bikes with shaped tubes, integrated headsets, and carbon seatstays: Why not integrate these features if it's commercially riskier not to? (And regardless of actual technical value, or lack thereof.)
In more recent years makers of tri bikes have mitigated their stand on wheel size, and that includes QR. Generally, bikes starting in the midrange size—say, 54cm to 57cm—will come in 700c, and bikes at or below that size will be available in 650c. This seems to be a sound compromise and explains in part why bikes are now creeping back toward 700c, as is demonstrated by the graph below.
There is also obviously an interest on behalf of those participating in the Ironman to buy bikes that offer the most features for the dollar. Carbon road race bikes dominate this category, specifically Trek's OCLV line. Kestrel also contributes to this with its Talon—the best-selling bike in its line. What drives this is the fact that molds cost a bike maker an awful lot of money to make. When you're going to spend between $50k and $100k for one size of one model, you can't afford to make them in triathlon geometry. There are not enough sales there. (Kestrel and Aegis are notable exceptions to this rule.) It's better, many of these companies think, to spend the money on molds that represent larger unit sales.
The question is, are triathletes buying OCLVs because of features-versus-cost or because they also desire a road-geometry bike? Or to put it another way, would Trek be more likely to sell a Hilo to a triathlete who'd otherwise buy an OCLV if the Hilo frame were constructed in the same way (and cost the same) as the road-geometry OCLV? Who knows?
One must also take into account the move, at least among age-groupers, away from steeper seat angles. If you're going to ride with road race geometry, you might as well ride a standard road race bike. Kestrel's 500 series is the only set of road bikes built with 650c wheels and slacker road angles. Virtually all others (with the notable exception of Trek's WSD road bikes built for women) come with 700c wheels. This begs the question, however, of whether OCLVs are chosen because of their road angles, or whether road angles are chosen because that's how the OCLVs come.
What "ought" the proper wheelsize breakdown be in Kona? To answer that question one must also tackle the question of what ought the seat angle breakdown be. Geometry is one of the determining factors in choosing wheelsize. Rider height is another. Comfort is another (700c wheels will, all things considered, give the rider a slightly more comfortable ride over the long haul of 112 miles). All that said, there is also the wild card of Kona's winds. Would a set of 650c large-surface-area wheels be easier to handle on a windy Kona day?