Seems an odd arrangement, nine boilers serving four sets of turbines, so the only way to develop full power would be to have them all operating as one unit. This of course raises the possibility of a hit or rupture anywhere in the system venting steam from the entire plant. In most large ships of the era, the number of boilers was a multiple of the number of engines, to allow split-plant operation. Or one could have an arrangement like the French Richelieu, six boilers and four shafts, so they could at least operate as two independent units. Even with all the boilers concentrated forward of the engine rooms, steam lines could be split by closing the appropriate valves. One might argue that with nine boilers, there would often be one down for routine maintenance, but on the day of battle one should be able to use all available power. In the WWI era and earlier, boilers were smaller, ships had as many as 42 and someone will probably cite an example of more; but by the 1930s new or reconstructed capital ships or cruisers had 4-12, serving 2, 3, or 4 shafts. I can't think of another apparent mismatch like the Yorktowns - anyone? - or a good reason for the arrangement. p.s. Many destroyers in various navies had three boilers for two shafts.
The decision made was to place the nine boilers in three transverse triplets ahead of the four engine rooms as a weight savings measure. Alternating engine and boiler rooms would have added additional transverse bulkheads, exhaust trunking, and longer propeller shafts, which would have added weight. It was felt that putting all the engine rooms and boiler rooms together was an acceptable risk, given that it was anticipated maximum torpedo warhead size would be c. 440 lbs of TORPEX. I am unclear however exactly how they plumbed the steam lines to the engine rooms and what arrangements for managing boilers there was. Friedman would be the go to for that information.
The Yorktown steam plant was the first in the US Navy to use superheat. The forward six three-drum boilers produced saturated steam and fed the aft three boilers which featured superheat and economizer sections. The aft three boilers then supplied superheated steam to the engine rooms. It was an interim design based on time and space limitations.
Not all boilers would be online at the same time unless they were expecting trouble, and maybe not then. Plans here. (Plate 14 applies.) Off the top of my career naval steam plant operator head I'd say the layout was to shorten the screw (propeller) shafts.
Thanks, gentlemen! So they had three "boiler units" supplying steam to four sets of turbines, and each superheat boiler had to be fed by one or two standard boilers. Well as you say, it was an interim arrangement.
A good discussion of the loss of CV-5 related to Ballard's 1998 expedition to locate her. Also discusses design considerations. http://www.sname.org/HigherLogic/System/DownloadDocumentFile.ashx?DocumentFileKey=1fc32bdf-e09c-4df3-affb-98f5c21772d9