Admittedly, I don't seriously practice figure skating, let alone do 3-3's

, but I don't think that underrotating the first jump would make completing the second one any easier. On the contrary, I think it would make the second jump much more difficult, because when a skater underrotates, he/she looses a lot of speed an flow out of the jump.
I like to create skating montages in my free time, which means that I had spent a lot of time looking at skater's jumps frame-by-frame. It seems to me that jumping 3-3's is the easiest for those skaters who are able to maintain perfect flow and speed in their jumps.
I've also noticed that skaters who just attempt to learn 3-3's usually choose as the first jump the triple which they can do the best and maintain the most flow out of. Usually it's a triple toe, but for example for Nana Takeda a loop jump is the easiest, so she atempted 3L-3T (and landed it in a show, although 3T was underrotated).
There are some skaters who have managed to eek out a prerotated triple on the end of an underrotated one, but their combinations didn't look pretty or easy, (and would get downgraded under current rules). In general I've rarely seen a skater underrotating the first jump, usually it's the second one that's more likely to be short on rotations.
If you do, say, a triple Lutz-triple toe, or triple flip/triple loop, is there a different landing technique on the first jump than when it is done solo?
I think it depends on the skater. I think that all the skaters would like a perfect edge and flow out of the first jump, but not all are able to do so yet. Some have managed to tack a triple, usually toe, on a mildly UR triple (like Kimmie), but I think that if they were able to sustain more speed out of the first jump and maintain more edge control, executing the second triple would be even easier for them.
If you'll look at this 3Lz-3T (0:46):
http://www.youtube.com/watch?v=ypCBsAZxIFo
it has the same flowy edge out of the first jump as the solo 3A, and both jumps are executed in the same way as if they were solo triples, fully rotated while still in the air. If one split the combo into two different videos, it would be difficult to tell whether they are a part of the combo or solo triples.
In the LP commentary at the same competition Dick Button mentioned that for a perfect combo "the first jump must be in perfect condition". I agree with that, and I think that the more control the skater has over the landing edge, the easier the second jump becomes.
Joesitz has made the point that a loop as the second jump in a combination is usually pre-rotated a lot more than a solo loop.
In general, I'd hesitate to use a term "pre-rotated" in relation to the loop jumps, because it suggests to me that a perfect loop should consist of precisely 3 rotations and the fact that the skaters "pre-rotate" it makes it somewhat easier. In fact loop seems to be a nemesis jump for many skaters and very few attempt 3-3L or even 3-2L combinations. It might not have
precisely 3 revolutions on the air, but this is how the jump mechanics work, and it's not any less"pre-rotated" that 3A is "over-rotated".
As for whether or not a loop as the second jump in a combination is usually pre-rotated a lot more than a solo loop, I think this is not always the case:
http://www.youtube.com/watch?v=h95n83KoBTY
This is a quick montage comparing Nobunari's loop as a part of 3Lz-3L combo with solo 3L out of running 3s and out of a normal entrance. All 3Ls are executed in the same way.
It would be interesting to compare combo loops and toes with solo loops and toes of other skaters, but I run out of time today.
