The real point to strike one in the matter is that the Empress should have landed in Sussex when her friends were awaiting her in the west—for Mr. Howlett fails to realize that she trusted to them and not to an "army" of her own.[835] The most probable explanation, doubtless, is that she hoped to evade Stephen, while he was carefully guarding the roads leading from the south-western coast to Gloucester and Bristol. Robert of Torigny distinctly implies that Stephen had effectually closed the other ports ("Appulerunt itaque apud Harundel, quia tunc alium portum non habebant").
In any case Mr. Howlett's endeavour to harmonize the two conflicting dates—the end of July and the end of September—by suggesting as a compromise the end of August, cannot be pronounced a success.[836]
It may afford, perhaps, some fresh light if we trace the king's movements after the arrival of the Empress.
Though the narratives of the chroniclers for the period between the landing of the Empress and the close of 1139 are at first sight difficult to reconcile, and, in any case, hard to understand, it is possible to unravel the sequence of events by a careful collation of their respective versions, aided by study of the topography and of other relative considerations.
On the landing of the Empress, the Earl of Gloucester, leaving her at Arundel, proceeded to Bristol (_Will. Malms._, p. 725). Stephen, who, says Florence's Continuator (p. 117), was then besieging Marlborough, endeavoured to intercept him (_Gesta_, p. 56), but, failing in this, returned to besiege the Empress at Arundel (_ibid._; _Cont. Flor. Wig._, p. 117; _Gervase_, i. 110). Desisting, however, from this siege, he allowed her to set out for Bristol.[837] Meanwhile, her brother, on his way to Bristol, had held a meeting with Brian fitz Count (_Will. Malms._, p. 725), and had evidently arranged with him a concerted plan of action (it must be remembered that they intended immediate revolt, for they had promised the Empress possession of her realm within a few months[838]). Brian had, accordingly, returned to Wallingford, and declared at once for the Empress (_Gesta_, p. 58). Stephen now marched against him, but either by the advice of his followers (_ibid._) or from impatience at the tedium of the siege,[839] again abandoned his undertaking, and leaving a detachment to blockade Brian (_Cont. Flor. Wig._, p. 118), marched west, himself, to strike at the centre of the revolt. He first attacked and captured Cerney (near Cirencester), a small fortress of Miles of Gloucester (_Gesta_, p. 59; _Will. Malms._, p. 726), and was then called south to Malmesbury by the news that Robert fitz Hubert had surprised it (on the 7th of October) and expelled his garrison (_Will. Malms._, p. 726; _Cont. Flor. Wig._, p. 119; _Gesta_, p. 59). Recovering the castle, within a fortnight of its capture (_Will. Malms._, p. 726), after besieging it eight days (_Cont. Flor. Wig._, p. 125), he was then decoyed still further south by the news that Humphrey de Bohun, at the instigation of Miles, had garrisoned Trowbridge against him. Here, however, he was not so fortunate (_Will. Malms._, p. 726; _Gesta_, p. 59). In the meanwhile Miles of Gloucester, with the instinct of a born warrior, had seized the opportunity thus afforded him, and, striking out boldly from his stronghold at Gloucester, marched to the relief of Brian fitz Count. Bursting by night on the blockading force, he scattered them in all directions, and returned in triumph to Gloucester (_Gesta_, p. 60). It was probably the tidings of this disaster (though the fact is not so stated) that induced Stephen to abandon his unsuccessful siege of Trowbridge, and retrace his steps to the Thames valley (_ibid._, pp. 61, 62). This must have been early in November.[840]