I am sure that 'Mr. S' is correct in his observations. It is really not that uncommon to find that projectiles of closely related types and of apparently identical external geometries range somewhat differently both regarding range vs Q.E. per se, and regarding dispersion at the target. It is (or was) quite common for artillery types to deliberately manufacture shells with identical or nearly identical external geometries so as to attempt to avoid needing new range tables, etc. in the field. Sometimes it worked, sometimes it didn't.
With new projectiles, if the differences in ranging etc. are important and the budget sufficient, the causes of these differences can inevitably be 'hunted down' via the applications of new technologies. With older projectiles, this is rarely the case.
One will find, in general, that dispersion at the target at long range is primarily controlled by consistency of initial velocity and consistency in drag. It turns out, for any of a number of reasons, that some combinations of projectiles and propellant just result in a more consistent intitial velocity. In some cases, getting a small I.V. scatter is effortless, whilst in others, attaining this goal might prove nearly (or, in rare cases, actually) impossible. Getting this right remains somewhat of a 'black art', in which experience and intuition sometimes play a major part, albeit (usually) backed by a considerable background in theory and technology.
The sources of inconsistent flight dynamics, i.e. inconsistent or unpredictable in-flight drag, can be equally difficult to predict. Yaw is usually not an issue, as the variables that control yaw over the course of the trajectory are usually fairly stable over a relatively wide range of conditions, provided the mount itself is in good condition. My experience has been that in most cases, drag inconsistencies stem from some relatively small geometric issue that renders the transition point from laminar to turbulent flow a bit unstable, from irregularities in the configuration of the driving bands after ejection (e.g. 'fringing') from some instability of flow around the base, or from some often relatively small variation in geometry which affects the origin and initiation of the shock waves which eminate from various and sundry positions along the projectile during flight. Tracking all of this stuff down, especially in the '30s, was very difficult indeed.
Simple drag and/or initial velocity inconsistency usually simply results in a larger pattern. This is not necessarily bad, particularly if target location is relatively ambiguous. Quite often, if the instability is large and takes the form of an 'on-off' variation, the pattern not only becomes large but develops a 'hole' in the middle, which can be very frustrating accuracy-wise.
Often the effects of drag instability and/or initial velocity instability are intertwined, and it can -- or at least could, prior to ubiquitous radar velocimeters -- be very difficult to separate them. A lot can be done by careful analysis of the fall of shot patterns, as each pattern type contains, in effect, the 'fingerprint' of the problem(s) which created it. A bit like reading tea leaves or goat entrails sometimes...