First of all, development at Apple is no different than anywhere else.
The development lines of 10.3 and 10.4 diverge. Instead of fixing faulty 10.3 code, it might be easier to rewrite it from scratch.
The vast majority of code in an operating system is common from major release to release.
What you seem to be arguing is that if a bug were discovered in 10.3.x it would not be fixed in 10.4 and that is totally wrong.
In fact lets imagine a bug. A USB dongle does not work with 10.3.4 ( to pick a number rather than a .x). It turns out to be Apple's fault. The manufacturers contact Apple and describe the bug. An engineer is tasked to fix the bug for 10.3.5. The manufacturers can then release with a "works with 10.3.5 and later " badge on the website, or on the box of the shipping product.
Does the engineer:
1) fix it for 10.3.5 only and ignore the fix for 10.4 or
2) Fix it for both
If he does 1) he gets fired. It does not matter if he had to write totally different code to fix the problem in 10.4, by the way, he still must fix it, or all point fix bugs reappear again in 10.4!
In many many cases he would not have to do anything other than add the same code to the same file, which would prbably be common to both releases. If not that simple , however, he stil has to fix it in both trains.
The only case where this would not apply would be if a feature was missing. If a feature was radically changed it is not an excuse to introduce the bug in the next full release ( 10.4) and makes the external manufacturer's liars.
Seriously. An engineer to do that would be fired ( although questions owuld be asked of the testers too).