One of two suppositions therefore:--(1.) If Phillips is right in his statement that Milton's first Divorce Tract was caused by the obstinate refusal of his wife to return to him, and the insulting conduct of her family in detaining her and laughing at his letters and messages, then Phillips's dates in the whole matter of the marriage must be a little wrong. "About Whitsuntide it was (May 21, 1643) that my uncle left us in Aldersgate Street, on what turned out to be his marriage journey; in about a month's time he returned, bringing his wife, and some of her relations, with him (June 1643); the relations stayed about a week, during which there was much feasting and merriment; for about a month after they were gone the newly-married wife remained with my uncle; but then (late in July or early in August 1643), tired of a philosophical life, and pining for the society of home, she contrived a request from her family to have her with them during the rest of the summer--to which my uncle consented, on the understanding that she was to come hack about Michaelmas (Sept. 29, 1643)." Such, re-expressed in words for the nonce, is Phillips's account as we have already given it. But, as the Divorce Tract was published August 1, 1643, it is clear that, if the cause of that Tract was the persistent, protracted, and contemptuous absence of his wife, then Phillips's memory must have been at fault, and he must have somewhat post-dated the marriage itself. The marriage in that case must have been before Whitsuntide 1643; and the return of the wife to her relations, her refusal to come hack, and Milton's chagrin and anger so occasioned, must have been matters not of after Michaelmas 1643, but of at least a month or two before the August of that year. This is quite a tenable supposition; for there are other inaccuracies in Phillips, and the register of the place and date of Milton's marriage with Mary Powell has not been found. (2) On the whole, however, Phillips's recollections about the marriage are so circumstantial, and there is such a likelihood of their being true, that, until contradictory records shall be produced, it seems right to accept his dating. But then his explanation of the cause of his uncle's speculations about divorce must be wrong. The cause in that case cannot have been the obstinate refusal of his wife to return; for the Divorce Tract must have been written and ready for the press while she was still with him in the Aldersgate Street house (July 1643), and it was actually out (Aug. 1) before she can have reached her father's house at Forest Hill on her granted two months of leave till Michaelmas. What are we to make of this discrepancy? One is puzzled. That a man should have occupied himself on a Tract on Divorce ere his honeymoon was well over--should have written it perseveringly day after day within sound of his newly-wedded wife's footsteps and the very rustle of her dress on the stairs or in the neighbouring room--is a notion all