Desktop Search. I tried a page that's completely roboted-out. On-Demand and Mobile Search came through loud and clear; Desktop Search is missing.
IP 220.127.116.11-99 (random)
UA for both: Mozilla/5.0 (X11; U; Linux x86_64; en-US) AppleWebKit/534.51 (KHTML, like Gecko) Chrome/12.0.742 Safari/534.51
referer for both: none for html, otherwise same as human
fetch order: html, css, js first, followed by all images
One copy of the page html. One copy of the shared stylesheet (accessible to all robots). Single copies of two (apparently random) images. Two copies of everything else including matching 403's on piwik.js.
If I saw this in logs I would never know it wasn't human, except for the missing favicon (only mobiles and repeaters skip it in real life).