Bearing this in mind we find that in the edition of 1633 there are only two poems--Basse's _Epitaph on Shakespeare_ and the _Psalme 137_, both already mentioned--for the genuineness of which there is not strong evidence, internal and external. But these two poems are the _only_ ones not contained in _D_, _H49_, _Lec_ or in _A18_, _N_, _TC_. In _D_, _H49_, _Lec_, on the other hand, there are no poems which are not, on the same evidence, genuine. There are, however, some which are not in _1633_, seven in all. But of these, five are the _Elegies_ which, we have seen above, the editor of _1633_ was prohibited from printing. The others are the _Lecture upon the Shadow_ (why omitted in _1633_ I cannot say) and the lines 'My fortune and my choice'. There are poems in _1633_ which are not in_ D_, _H49_, _Lec_. These, with the exception of poems previously printed, as the _Anniversaries_ and the _Elegie on Prince Henry_, are all in _A18_, _N_, _TC_. This last collection does contain some twelve poems not by Donne, but of these the majority are found only in _N_ and _TCD_, and they make no pretence to be Donne's. Three are initialled 'J. R.' (in _TCD_), and two of these, with some poems by Overbury and Beaumont, are not part of the Donne collection but are added at the end. Another poem is initialled 'R. Cor.' The only poems which are included among Donne's poems as though by him are _The Paradox_ ('Whoso terms Love a fire') and the Letter or Elegy, 'Madam soe may my verses pleasing be.' Of these, the first is in all four manuscripts, the second only in _N_ and _TCD_. Neither is in _D_, _H49_, _Lec_, or _1633_. The last is by Beaumont, and follows immediately a letter by Donne to the same lady, the Countess of Bedford. Doubtless the two poems have come from some collection in which they were transcribed together, ultimately from a commonplace-book of the Countess herself. The former _may_ be by Donne, but has probably adhered for a like reason to his paradox, 'No lover saith' (p. 302), which immediately precedes it. We have thus three collections, each of which has kept its canon pure or very nearly so, and in which any mistake by one is checked by the absence of the poem in the other two. It cannot be by accident that these collections are so free from the unauthentic poems which other manuscripts associate with Donne's. Those who prepared them must have known what they were about. Marriot must have had some help in securing a text on the whole so accurate as that of _1633_, and in avoiding spurious poems on the whole so well. When that guidance was withdrawn he was only too willing to go a-gathering what would swell the compass of his volume. If then a poem does not occur in any of these collections it is not necessarily unauthentic, but as no such poem has anything like the wide support of the manuscripts that these have, it should present its credentials, and approve its authenticity on internal grounds if external are not available.