Giuliano Colliva's "Questioni di tiro... e altre", Bollettino d'Archivio dell'Ufficio Storico della Marina, settembre 2003, dicembre 2003 and marzo 2004, debunks several alleged causes that have been put forward about the alleged dispersions of Italian guns, first of all the variability in the weights of shells and charges. The author also notes that in the battle of Cape Spada the CA Sydney shot 956 shells scoring only three hits.
Apologies to respond to this months after it was posted, but would you perhaps know where one might get their hands on these Bollettini? Unfortunately those available online on the website of the MMI only go back as far as 2011.
---
That aside, it's worth noting that there's certainly been a turn against the historiography of 'shell quality control variability' over the last few years. The 2018 Storia Militare 'Briefing' on the
Zara-class cruisers, by Maurizio Brescia and Augusto De Toro, directly states that the period of variable quality control only lasted up to 1937, and that the actual reduced tolerance was 0.4 to 0.5%, rather than the 1% commonly cited in the past. This is part of a larger discussion on the RM's investigation into dispersion issues noted on certain calibers of guns just before the war (in particular the 203/50 and 203/53, but also the 152mm and 320mm guns), which actually established the primary issue as being errors produced by the Follow-the-Pointer system used to transmit elevation and train data to the gun turrets caused by 'drag' in the electrical systems, and 'slack' in the transmissions of the the gun movements back to the fire control system. The 'drag' was found to produce errors in gunlaying, though this was less so the case on the newer systems used on
Conte di Cavour and
Giulio Cesare.
Overall the report of the commission recommended or stated the enactment of various fixes or further investigations relating to these issues, though it isn't clear how many would have actually be conducted in practice. The report is dated 13 May 1940, and of course Italy was soon at war.
Bagnasco & de Toro also touch upon the dispersion issue in their pair of Storia Militare 'Dossier' from 2020 on the
Conte di Cavour and
Duilio-classes (also combined into a single volume in English), where they note the steady process of perfecting the at first very troublesome 320/44's, including various efforts to combat dispersion. Ultimately, the issue could not be solved on these guns - they seem to have just naturally not had tight dispersion - but in the process of doing this threw out a wide range of potential issues as they were investigated, including issues in the FTP systems (gunlaying errors on these newer systems were consistently less than a thousandth), variability in propellant charge weight or burn rate, issues with shell shape, driving bands, etc. I would also note that incorrect behavior in flight on the part of the projectiles was on occasion an issue for individual shells, which seems interesting to note given the stated conclusions of the Warships International article cited above (unfortunately I don't have access to the article so I'm not sure if that's addressed in it or not, I only know what I've been told about the article).
At the end of the section discussing the gun dispersion up until the start of the war, Bagnasco & de Toro directly state that in all the documentation the consulted from the late 1930s and from the war itself, there was
no mention of variations in projectile and charge weights attributed to excessive tolerances accepted in ammunition supply - though the aforementioned concern about dissimilar combustion rate of charges did exist as a concern for charges. They also reject the less popular but still sometimes mentioned idea that barrel droop might have contributed to the gun dispersion. Overall, on the dispersion issue, they state (I quote from how it is translated in the English-language book)
"This does not mean that the problem did not exist, as the Regia Marina hierarchy was aware of it and sought to remedy it, but in an absolute sense, it was not as serious as many would have liked it to be believed, especially in compared with the performance of other big guns such as the 330/50 and 380/45 of the 'Dunkerque' and 'Richelieu' class battleships."
In regards to speed, I always understood that the RM rarely conducted trials in a full-load condition. The recorded speeds were never seen in actual operations. Speed dropped due to the increased weight of modifications. I don't know if the machinery was already worn-out by the time the war had started. If it was, then the machinery wasn't very robust, also indicating that the ships were light-built to save weight.
Apologies also for responding to something two months after if was posted, but in regards to speed;
There are multiple types of trials a ship can embark on, and in the RM's case, the most famous of these were the machinery trials intended to push propulsion plants to their limits, with speeds over contract value resulting in bonuses to builders. These speeds were good for propaganda figures, but these were also certainly different to actual service speed trials and such figures were not used by the RM. Vincent P. O'Hara and Enrico Cernuschi have an brief article/note in Conway's
Warship 2006 that discusses this in regards to the
Giussano-class large scouts/light cruisers, noting at the top speed credited to the class by the ship's handbook was 36 knots when the ships were brand new, and by 1939 they were credited with a top speed of no more than 33 knots.
In terms of designs speeds being met, this somewhat depending on the design. Destroyers very rarely met the 38 knot designed top speed that was standard for the 1930s, since weight impacts from modifications were proportionately much greater on them then on cruisers, and their much more active careers tended to cause greater wear on them then other ship types - thus most Italian '38-knot' destroyers tended to have a maximum service speed closer to 32-35 knots (depending on the class). The rebuilt Navigatori (essentially all but
Nicoloso da Recco and
Antonio Usidomare) meanwhile were ostensibly 34-knot destroyers after their reconstruction but more like 28 knots in service. Cruisers, excepting the
Giussano and
Cadorna-classes, tended to be better, though the
Trento-class struggled to output the intended 150,000 shp and so tended to be more like 33-34 knot ships rather than 35 knot ships.
This is not exclusively an Italian issue, particularly in regards to the destroyers - I would note that among French destroyers of the 1920s also generally did not reach their designed top speed. The
Jaguar-class were 31-knot ships rather than their designed 35.5 knots, while the
Bourrasque and
L'Adriot-class tended to be 29-knot ships versus their 33-knot design speed. Later classes seemed to have largely overcome this issue, though these are notably considerably larger ships than even the Navigatori-class light scouts (except for the
Le Hardi-class), and could much more easily fit very powerful propulsion plants.