Rossetti most likely began writing “The Burden of Nineveh” in October or November of 1850, shortly after the enormous Assyrian bull bas-relief arrived in London.@On the poem's composition, see The P.R.B. Journal, 82; and W.M. Rossetti's Dante Gabriel Rossetti as Designer and Writer, 137-38 (cited from the Rossetti Archive). The story of the monument's arrival in London can be found in Bohrer, 125ff. See also Malley, Jenkins, 155ff.; and Gadd, 124-27 and 159-60. It has been thought that no record of the poem at this earliest stage survives. However, as a search in the Rossetti Archive reveals, we do have two stanzas (numbered 5 and 6) of what appears to be an 1850 fair-copy manuscript (now lost), preserved in facsimile in H.C. Marillier's Dante Gabriel Rossetti, An Illustrated Memorial of his Life and Art  (1899; cited from the Rossetti Archive). In addition, Marillier reproduces a second manuscript fragment (of stanza 7), which he dates somewhat later, but prior to the poem’s first publication in 1856 in the Oxford and Cambridge Magazine, the point at which scholars have thought the trail begins. These texts, our earliest witnesses to the poem, were recorded only through a stroke of luck: Marillier chose to reproduce pages from the two manuscripts of “The Burden of Nineveh” as examples of the changes in Rossetti’s handwriting between 1850 and 1856. He writes,

The arbitrary choice of these pages has led to their survival, since this particular manuscript book of Fairfax Murray’s is untraced.@Perhaps they were once part of the notebook headed “Lyrics &c.” now in the Troxell collection: http://www.rossettiarchive.org/docs/lyrics.prinms.radheader.html. In any case, Marillier suggests that they were part of a book that Fairfax Murray had assembled himself, in a format similar to “Lyrics &c.” Whether Marillier removed them from the book in order to make his reproductions is likely but uncertain. Perhaps this ‘stroke of luck’ was in fact from a double-edged sword, since the loss of the notebook may have stemmed from Marillier’s handling of it. The dating and relationship of the two specimens is presumably conjectural (Marillier’s third example is a manuscript of “Adieu,” written circa 1880). As Marillier indicates, the “Burden of Nineveh” manuscripts offer early examples of Rossetti’s handwriting, and it seems that these manuscripts represent comparatively early stages in the poem’s evolution (prior to its 1856 publication), since the readings they contain do not recur in any of the later texts.

    The verses Marillier presents from the 1850 fair-copy are free of revisions except for three corrections that must be original to the manuscript’s transcription (i.e., they are not later alterations, but running corrections of the lines). The following is the text of the first Marillier manuscript (see figure 1):

    As I have mentioned and will discuss shortly, these 1850 readings were substantially altered prior to the poem’s first publication. In addition, the stanza numbering is off: these lines are closest in content to stanzas 3 and 4 of the Oxford and Cambridge Magazine version (not 5 and 6), indicating a cancellation in manuscript. Furthermore, Rossetti added a stanza during the 1869 revisions – resuscitating the one deleted prior to 1856 and working from a manuscript very like the one that Marillier saw – so these numbers do square with those of the received text after 1870. That is, the apparently-belated 1869 addition is in fact a reinsertion of lines from the very early 1850 version. In the 1869 Penkill proofs, Rossetti added the following verses as the new 4th stanza, carried forward virtually unchanged into the first edition of the 1870 Poems:

    The second half of this stanza (“Oh seemed it not…”) was already present in the 1856 Oxford/Cambridge version, where it was fused with the first half of stanza 5 in the newly-discovered 1850 Marillier MS. This is the Oxford/Cambridge stanza (cited from the Rossetti Archive):

In other words, in preparing the poem for its first publication in 1856, Rossetti combined stanzas 4 and 5 from the 1850 version, judiciously cutting the clichéd and anachronistic section on “Heaven’s thunder,” “the cry of shells” and “lordships loud with bells…in secret Nineveh.” He also eliminated the dangling modifier in lines 2-3; “Blue and huge” seems meant to describe the shadow, not the “beast” itself.