ABSTRACT
Date and circumstances of composition. This poem’s remarkably intense period of composition — ‘three days’ — is recorded in l. 36 of its dedicatory stanzas to Mary. Its date may be identified precisely as 14–16 August 1820 from Mary’s journal (Mary Jnl i 329), where she records ‘W. W. A.’, likely an abbreviation for ‘write’ or ‘wrote’ ‘Witch of Atlas’, on the first two of these days and ‘Do — Finished’ on the last. S. had returned to Bagni di San Giuliano from a solitary two-day walk to Monte San Pellegrino, a peak in the Apennines in the territory of Modena, on the evening of Sunday, 13 August. Two days previously, during a period of ‘warm and delightful’ weather which made ‘[t]he Country … delicious’, he, Mary, and Claire had made the brief journey to Lucca where they spent the night, the latter pair remaining there the next day to look at sites relating to the historical basis of the novel Mary had begun writing, Valperga (1823), while S. made his weekend excursion (Mary Jnl i 328–9, Claire Jnl 169–70). As GM notes, WA has the flavour of ‘a holiday-poem’. The restorative effects of the recent move to Bagni di San Giuliano on S.’s health, and the felicitous Tuscan climate and countryside in the summer months that inform the poem’s energy and spirit, are captured in Mary’s recollection of the atmosphere in which it was written:
We spent the summer at the baths of San Giuliano, four miles from Pisa. These baths were of great use to Shelley in soothing his nervous instability. We made several excursions in the neighbourhood. The country around is fertile; and diversified and rendered picturesque by ranges of near hills and more distant mountains. The peasantry are a handsome intelligent race, and there was a gladsome sunny heaven spread over us, that rendered home and every scene we visited cheerful and bright. During some of the hottest days of August, Shelley made a solitary journey on foot to the summit of Monte San Pelegrino — a mountain of some height, on the top of which there is a chapel, the object, during certain days in the year, of many pilgrimages. The excursion delighted him while it lasted, though he exerted himself too much, and the effect was considerable lassitude and weakness on his return. During the expedition he conceived the idea and wrote, in the three days immediately succeeding to his return, the Witch of Atlas.
(1839 iv 50–1) WA is founded in several kinds of dialogue between S. and Mary, most overtly in its Dedication to her concerning the audience of his poetry. But as significant to both of them may be the place which prompted S.’s poem. Two summers before, on 2 July 1818, soon after they had made an excursion to the celebrated Apennine sight, the ‘Prato fiorito’ (the flowery meadow), from Bagni di Lucca (L ii 20), Mary told Maria Gisborne that ‘Mr Shelley wishes to go with me to Monte Pelerino — the highest of the Appenines at the top of which there is a shrine — It is distant about 22 miles — Can it be there that the Italian palates were deceived by unwholesome food (to talk of that hideous transaction in their own cool way) — ? and would you think it advisable for us to make this pilgrimage? — we must go on horseback and sleep in one of the houses on the mountain.’ (Mary L i 74) S.’s idea for this expedition to Monte San Pellegrino (literally ‘Mount St Pilgrim’ or ‘Mount St Wayfarer’) may have been originally prompted by an enthusiasm for following the Serchio towards its source, in accordance with earlier journeys to the sources of rivers such as the Thames in 1815, since a stream flowing from near the peak joins the Serchio in the valley below. But after his stay at Este, near Petrarch’s house at Arquà (L ii 43), in September–October 1818, S. may have modelled his walk on that poet’s solitary expedition to the summit of Mont Ventoux ‘to divert his thoughts’, as Mary later put it (MSLL i 21). In addition, the resumption of work on her novel after the move to Bagni di San Giuliano on 5 August 1820 raises the question of whether a new pretext for S.’s excursion was to undertake a research visit on Mary’s behalf. According to a medieval tradition recorded in Valperga, the source of which Mary’s working notebook Bod. [Abinger] Dep. e. 274, p. 95 shows to be Lodovico Muratori, Dissertazioni sopra le Antichità Italiane, 3 vols (1765–6) iii 210, ‘The san Pellegrino of Monte San Pelegrino was a king of Scotland [who] for love of God renouncing his kingdom made pilgrimages to various holy places and died on that mountain’. Annual expeditions to the peak of the mountain took place in the summer:This was the season of pilgrimages to Monte San Pelegrino, a wild and high Apennine in the neighbourhood of Valperga. It is said that a king of Scotland, resigning his crown to his son, and exiling himself from his country, finished his days in penitence and prayer on this mountain. In Italy every unknown pilgrim was a king or prince: but this was a strange tradition; and it would seem as if the royal penitent, disdaining the gladsome plains of Italy, sought for the image of his native country on this naked peak among the heaped masses of the Apennines.
His memory was there canonized, and many indulgences were the reward of three successive visits to his rocky tomb; every year numberless pilgrims flocked, and still continue to flock thither. Straining up the rugged paths of the mountain, careless of the burning sun, they walk on, shadowed by their broad pilgrim’s hats, repeating their pater-nosters, and thus, by the toil of the body, buy indulgence for the soul’s idleness.
(MSW iii 185–6) Even embarking from Bagni di Lucca to ascend the 1529m summit on foot would have meant a strenuous expedition, but starting from Lucca, further south, would have made it virtually impossible. S.’s subsequent exhaustion, remembered by Mary in the passage cited, suggests he had over-exerted himself notwithstanding his singular method of walking recalled by Hogg: ‘it was his delight to strike boldly into the fields, to cross the country daringly on foot, as is usual with sportsmen in shooting; to perform, as it were, a pedestrian steeple-chase.’ (Hogg i 110) GM notes of S. in 1817 that as ‘[a] powerful walker, he would cover the 64 miles from Marlow to London and back in two days, spending one active day in town in between.’ Because it is difficult to believe that even S. could have walked all the way to the peak and back in less than 48 hours, given that ‘these wooded mountains and deep river valleys would be almost impenetrable’ (GM), it is almost certain that he rode part of the way, and whether he ever reached the summit is unknown. While the main body of the poem must have been conceived during his excursion, given the position of the rough draft of the dedicatory stanzas in Nbk 14, Carlene Adamson has suggested that they were possibly written before, not after, the main body of the poem had been drafted (BSM v pp. xl — xli). In support of her hypothesis, Mary’s reaction to his plan to write the poem recalled in the Dedication may have occurred in a conversation during her ‘walk with S.’ which took place the day after his return (Mary Jnl i 329), that is, before he began his draft of WA. Mary later explained the Dedication thus:The surpassing excellence of the Cenci had made me greatly desire that Shelley should increase his popularity, by adopting subjects that would more suit the popular taste, than a poem conceived in the abstract and dreamy spirit of the Witch of Atlas. It was not only that I wished him to acquire popularity as redounding to his fame; but I believed that he would obtain a greater mastery over his own powers, and greater happiness in his mind, if public applause crowned his endeavours. The few stanzas that precede the poem were addressed to me on my representing these ideas to him.
(1839 iv 51) MS evidence and circumstances of publication. The surviving leaves of Nbk 14 contain rough draft of a substantial proportion of the poem, including the Dedication and 58 of the 78 stanzas, in whole, or in part, on pp. 94–70, 68–66, 30, and 9–8, all reverso. Carlene Adamson (BSM v p. xli) records the stanzas for which there is no draft as 1–4, 8–9, 12, 18–19, 22–24, 27, 29–30, 32–33, 37, 58 and 76 (there is a complete list of the extant draft keyed to line references in BSM xxiii 258). Draft of at least some of the missing stanzas is likely to have been on pages subsequently removed from the nbk. Adamson’s suggestion that the dedicatory stanzas may have been entered in Nbk 14 before the beginning of the poem is qualified by the fact that the progress of the rough draft through the nbk does not correspond to the order of the stanzas in S.’s intermediate fair copy of the dedication and main body of the poem in Nbk 16 ff. 15r — 16r and 17r — 32v. This copy, clearly based on Nbk 14, contains corrections in S.’s hand, some in darker ink, therefore done later, some lacunae (see notes to l. 96 and ll. 133–6), and some pencilled figures on f. 32v, including a line-count for the main body of the poem. Given its position in Nbk 16 soon after the intermediate fair copy of Ode to Naples (Longman iii 625–49, no. 343), this intermediate fair copy of WA was probably made in, or soon after, early September 1820 (see headnote to no. 343) and by early December at the latest. (Forman mistakenly refers to Nbk 16, which he never saw, as ‘a perfect MS.’ of WA in Forman 1876–7 iii 242). Both of the fair copies of the poem known to have been made by Mary are lost. It is possible that they were done in December 1820–January 1821. Mary records transcribing WA on 12 December (Mary Jnl i 342) and the further copying for S. on 19 December that was completed on 6 January 1821 (Mary Jnl i 343, i 348), a period during which he could not read nor write because of ophthalmia, may also have been of WA. A fair copy was sent to Ollier by S. on 20 January 1821, with the comment that it was ‘a fanciful poem, which, if its merit be measured by the labour which it cost, is worth nothing’ (L ii 257). A month later, on 22 February, he instructed Ollier explicitly, and for obvious reasons, not to publish it with the pamphlet containing J&M and the other melancholy poems referred to in his letter of 10 November 1820 (see headnote to no. 356, Ye hasten to the [grave]! What seek ye there, Longman iii 725–7) but invited him to ‘put my name to the “Witch of Atlas”, as usual.’ (L ii 269) Since Ollier did not act, a year later, on 25 January 1822, S. wrote to Hunt referring to WA as amongst ‘a parcel of little Poems’ which he sought Hunt’s assistance in trying to place (L ii 381). When Mary began to assemble 1824 for publication, she asked Hunt in a letter of 18 September 1823 to return her copy of WA: ‘You have I think my copies of the Essay on Devils — Translation of Cyprian — Witch of Atlas’, adding, with apparent reference to Nbk 16, ‘The Witch of Atlas I have a copy of corrected by S. — and those vacancies filled up which are in my copy.’ (Mary L i 384) From this letter, and another of 12 January 1823, informing Jane Williams that she expected WA to appear in the ‘next’ (Mary L i 307), i.e. the second, number of The Liberal (in fact published on 1 January but not seen by her until ‘a short time’ before 10 May (Mary L i 338)), it appears that she had given Hunt a transcript of the poem at Albaro in late 1822. This would have been as part of her plan to ‘publish my Shelley’s Mss. in the Liberal. First — being out of England & Ollier behaving so ill it is almost impossible to do it in any other way. — I think also that they will do good to the work & that would best please him — Then when I am rich enough I will make an edition of all he has written — & his works thus appearing at intervals will keep him alive in the minds of his admirers.’ (Mary L i 306–7) WA’s irreligious, light-hearted, and ‘southern’ feel would have made it eminently suitable for that periodical. The transcript Mary gave to Hunt, identified as MT, is mentioned in Forman 1876–7 iii 242 as ‘among the Leigh Hunt MSS. placed at my disposal by Mr. Townshend Mayer’ after Hunt’s death. Forman records its ‘variations from the received text’ in Forman 1876–7, but it should be noted that many of the variants in MT correspond to those in Nbk 16 and those which do not are explicable as mistranscriptions of the kind Mary made in her transcription of LMG. As E. B. Murray suggests (BSM iv, Pt II, 336), most of its differences from 1824 appear to result from MT being based on an early version of the intermediate fair copy in Nbk 16, i.e. before all the corrections to it had been made by S. The variants recorded in the notes to ll. 519, 609 and 628 point to MT having begun as a safe-keeping copy made in December 1820–early January 1821. Because MT was not returned to her, the basis of 1824 is likely to have been, as Murray suggests (BSM iv, Pt I, p. xxx), her first fair copy that S. had sent to Ollier in January 1821, probably amongst the MSS she acknowledges having retrieved from Ollier in a letter of 19 November 1823 (Mary L i 402). The poem was published without the Dedication in 1824, with its first three stanzas in 1839 and with all six in 1840. Given the publication of PB3 in 1840 (see headnote to no. 239, Longman iii 70–81), there was now no justification for not publishing the Dedication’s last three stanzas, with their reference to Wordsworth’s Peter Bell. Based on his consultation of Nbk 16, Garnett published in Relics 94 corrections to readings in 1839 and 1840 of ll. 109, 210, 424, 596 and 599. Taylor 73 notes that 1834 appears to have been the source of three readings in 1839 not in 1824, those in ll. 210 and 599 corrected by Garnett, and another in l. 549. The only other reading in 1839 not in 1824, in l. 333, occurs in 1829.