Mediterranean Fleet cruiser commander Rear Admiral B Wake-Walker (who would later write the Royal Navy’s cruiser handbook) pointed to the great strategic value of speed demonstrated by recent manoeuvres. ‘It is not sufficiently realised that at sea as on land, to have the advantage of speed is to strategically double one’s value.’ A speed advantage was like the weather gauge sought by seamen under sail: the fleet possessing it could bring the enemy to action and prevent him from getting away. Only when fleets were in contact and both anxious to fight was speed less important. ‘But what is the good of a perfect battle fleet that can never hope to overtake the enemy and unless it is able to drive them into a corner, cannot force an action?’ He was well aware of the argument that if the Royal Navy built for 18 knots, some other country would build for 18.5 knots and so on. Wake-Walker’s solution was to build a special class of ships specifically to ‘hang onto the rear of the enemy’s fleet’. He imagined 14,000-ton ships with 2500 tons of coal and a speed of 22 or 23 knots, a complete armoured belt (specially designed so they would not lose speed to bow hits) and bow and stern fire of three or four 9.2in guns, plus a battery of ten or twelve 7.5in guns. Three such ships, ‘judiciously handled’, would be a match for two battleships. Their attacks on the rear of the enemy line would slow enough ships to slow down the entire enemy fleet. These ships might be called second class battleships or fleet armoured cruisers. In the recent manoeuvres a Duncan class battleship was assigned to stiffen the cruisers. This policy proved successful, so Wake-Walker advocated a unit consisting of a Duncan and two Cressy class armoured cruisers. Moreover, if an enemy attached armoured cruisers to its fleet to deal with British armoured cruisers, they would not be available to raid British commerce. ‘We thus force them into a policy, in spite of themselves, in other words we start the war holding most of the trumps in our hands.’
The conning tower is clearly visible at the fore end of Agamemnon’s superstructure.
The three 1903–4 ships were not originally to have been repeat King Edward VIIs.9 Philip Watts was now DNC: the previous year he had replaced Sir William White. Much as White had replaced Barnaby as a more forward-thinking cruiser designer from private industry (Elswick), it seems that Watts was considered a more forward-thinking battleship designer from the same company. He was well acquainted with Admiral Fisher, who when Controller had assured him that he would become DNC. Watts began by claiming that he could have provided the characteristics of the King Edward VII class on 14,000 tons rather than 16,350 tons. The difference could go into heavier protection or more weapons or both. Watts offered six outline designs for 14,000-ton battleships. The repeat King Edward VIIs were built because in August 1903 the Board cancelled Watts’ new design.10
Watts began work on the next battleship in February 1902. He produced an outline of a 14,000-ton ship with the same armament and protection as the King Edward VII class. In April he was asked to enlarge the ship to carry four more 9.2in guns (total of eight). In May he reported the size and cost of an enlarged ship carrying four 12in, twelve 9.2in and twelve 6in guns, with a speed of 18.5 knots: 19,000 tons, £1,700,000. This was far too large for the Board. The alternative was an enlarged version of the last pre-King Edward VII design, the Duncan class: it would retain a speed of 19 knots and the same protection (considerably thinner than a King Edward VII), but would have reduced stores. Armament would be four 12in, eight 9.2in and ten 6in. That would require a ship of 15,265 tons or 15,880 tons with full stores (5 July 1902).
Watts began work on a new design incorporating his ideas of greater efficiency. On 8 July 1902 he reported four designs (A to D) of a 14,000-ton ship with ‘special arrangement of armour’ and various speeds and gun arrangements, with armour fixed by what was left over within the 14,000-ton limit.11 In effect he was paring his ship back to the displacement of the earlier Duncan, which had a conventional battery (four 12in, twelve 6in) and considerably less protection than a King Edward VII (7in side and no upper belt), but did offer higher speed (19 knots vs 18.5 knots). Watts’ new approach made it possible to build much more heavily-armed ships within the limits the Royal Navy had to accept. The May report showed, in effect, that what became the Lord Nelson class, the bridge to the Dreadnought, could not be built using White’s design practices.
Watts’ Legend for the initial series gives some idea of how he hoped to save weight. Design A offered the same armament as a King Edward VII and the increased speed (19 knots rather than 18.5 knots) of a Duncan. Compared to King Edward VII, Design A required 190 tons less general equipment, the greatest saving being in anchors and cables. Armament weight was somewhat reduced (2200 tons rather than 2575 tons), presumably by rearrangement. Machinery was cut from 1800 tons to 1430 tons, probably mainly by adopting lighter-weight boilers. Coal was cut from 950 tons to 880 tons. Armour and its backing were cut from 4175 tons to 3780 tons. Hull structural weight was cut from 5900 tons to 5170 tons, partly because A was a shorter ship (410ft rather than 425ft between perpendiculars). It offered about the same beam, but slightly less draught. A smaller ship required less power for the same speed: 15,000 IHP rather than 18,000 IHP with natural draught. The 14,000 tons included no Board Margin (in King Edward VII it was 200 tons).
Lord Nelson shows her charthouse, with a compass platform above and a flying bridge well above that. Later ships eliminated the flying bridge and were conned from the compass platform above the charthouse (which contained the helm, with another helm in the conning tower). The 12pdr battery was mainly on the flying deck above the 9.2in turrets, with 3pdrs atop the turrets, as in the King Edward VII class.
In the summer of 1902 the Board accepted Watts’ Design A as the basis of the 1903–4 programme. He was asked to try a variety of alternatives. Giving up the 6in battery altogether would save 290 tons, which could be invested in armour, for example in increasing citadel armour from 4in to 7in (A1); alternatively, some of it could be added to the bow (A2). The main deck amidships over the machinery could be thickened from 1½in and 2½in (flat/slope) to 2in and 3in (150 tons, version A3). Alternatively, giving up half a knot (reduction to 13,500 IHP) would save 140 tons, which might equate to a 5½in citadel (A4) or increased bow armour (A5) or a thicker deck amidships (A6).
Watts’ Design B offered eight 9.2in guns and twelve rather than ten (as on King Edward VII) 6in. Watts said it had the armament of the American battleship which had inspired the King Edward VII in the first place and the speed of a King Edward VII. To provide them within a fixed displacement, he drastically cut ammunition per gun: fifty-four rather than the usual eighty rounds per 12in gun, 100 rather than 150 per 9.2in, 166 rather than 200 per 6in, 200 rather than 300 per 12pdr and 333 rather than 500 per 3pdr.12 Watts retained the 9in waterline belt, backing it with a thicker armoured deck (2½in slope, 1½in flat) – but eliminating the upper armoured deck of the earlier ship. He narrowed the upper part of the belt and cut it to 7in. The upper belt, which now extended below the main deck, was cut to 4in (but with 7in over the batteries). Watts estimated that the resulting ship would cost £1,150,000 compared to £1,300,000 for a King Edward VII with a less powerful armament (however, Watts’ ship was much more expensive than a Duncan because of its more powerful armament; the earlier ship cost £980,000).
White had often argued that commercial builders sold warships to foreign navies on much the same basis Watt had designed his ship: more guns, fewer