A blaze of glory: The remarkable true story behind autumn’s greatest show of colour

Under constant attack by the elements... One vote away from the chainsaw... That Westonbirt – the national arboretum – survives, is down to the work of one man. David Randall pores over his diaries

This is the story of an Englishman who loved trees. He worked with them, lived among them, and kept a diary of their lives, each year recording their bursting into leaf, dense summer canopies, and winter's pinchings by frosts. As he walked the woods he would know for 60 years, no fallen bough escaped him. And when a tree died, he would return to his cottage, settle in his parlour, and write its obituary ("It is with much regret...") as if it were some venerable colleague who had passed away after a lifetime of loyal service. But mostly, in his neat, businesslike hand, ' he wrote in great detail of the climax of the trees' year: itemising and judging each autumn's display of colour as if it were entered in a show. And he would wrestle with the mystery of why trees blazed like a well-made fire some years, and yet, in others, merely glowed like embers.

His name was William John Mitchell, and he was the first curator of Westonbirt, now the national arboretum. In the next few weeks, more than 100,000 visitors will flock here for the finest autumn colours in all England. As they walk these avenues, made red and gold by the dying of the year, they should know of the man who loved these trees more than any other, of how they nearly broke his heart, and how close his woods once came to the chainsaw.

To really understand how special this place is, we must go back to the 1830s, and Westonbirt's founder Robert Stayner Holford, MP, art collector, and a man with a huge advantage for someone who had set their heart on remaking a landscape – immense wealth. At the age of 30, he inherited £1m (equivalent to £50m today), and, reportedly, a wheelbarrow full of gold as well. The source of these funds was a substantial shareholding in the New River Company, a 42-mile long canal which brought fresh water from Hertfordshire to London. A few years later, Holford inherited no fewer than six estates from his uncle. Thus, unencumbered by the need to watch pennies, pounds, or guineas, Robert settled into Westonbirt and set about creating his arboretum in this drowsy corner of Gloucestershire.

He looked, with his Cavalier moustache, beard, and broad-brimmed hat, like the villain in a Victorian melodrama. Yet Holford's passion was not the ravaging of maidens, but the planting of trees, and for this he had a good eye. He arranged his arboretum not in some cataloguists' way, with all the South American native trees here, and Japanese ones there, but with a landscape artist's knack of making views, mixing pleasing combinations with unexpected juxtapositions. Soon, hundreds of acres of downland began to fill with specimen trees from all over the world, and he would invite friends down to picnic among the autumn leaves at one of his "colour parties". Robert died, his son George Lindsay Holford greatly expanded the collection, and, on his death, Westonbirt passed to his nephew, the 4th Earl of Morley, a man with a keen interest, but not the bottomless pockets of the Holfords. The great house was sold for a girls' school (which it still, thrivingly, is), and a curator sought for the arboretum.

The man chosen was Mitchell. Born in Plympton, Devon, in 1876, he had gone to work at Westonbirt as a young man, and become foreman gardener by the age of 25, He was 51, married with five daughters, and the most senior horticulturalist on the estate. He now had charge of around 50 men and horses, and, to the task of superintending 600 acres of specimen trees, he added the keeping of seasonal logs. The existence of Mitchell's notebooks was, until recently, known only to his family, but they are now being transcribed for a wider audience. In them, between 1928 and 1947, he recorded each season's weather and its effects on Westonbirt's trees. Here, he wrote not only of the good times but the bad, such as the gales of 1930; the extraordinary frost of late May 1935 ("the Acer volxemi looked as if it had been sprayed all over with boiling water, not a live leaf remained"); the burning summer which followed ("During July, the temperature soared to over 80 degrees in the shade for 21 days... for six weeks following, two horses and water barrels, with the accompanying men, did nothing but haul water to the thirsty things"); and, especially, the ice storm of January 1940.

This was that rare phenomenon called a "glaze", where rain falls and immediately freezes on contact with any surface. AW Hughes, editor of the Dursley Gazette, reported "telephone wires made as thick as bell-ropes... birds frozen as they roosted", while Mitchell looked on, appalled, as the ice built up on his trees. He wrote: "Every twig and leaf was coated with ice from half an inch in thickness on the sheltered side to one-and-a-half inches on the side exposed to the SE to E wind. Eventually the tremendous weight of this accumulated ice was too much for the trees to bear, and with reports like explosions, the largest limbs cracked and fell. ... Every tree in the park had from a cartload to a wagon load of broken branches underneath it, and the majority of them were completely stripped one side. Others looked as though they had been beheaded, the terrible rents in others showed all too plainly the terrible ordeal they had suffered... It was a most fantastic sight, and when the ice-coated branches were moved by the wind, they made a weird sound like the rattling of chandeliers."

Each year's log is written like a story whose ending is the timing, quality and longevity of that autumn's colours. Every species, and sometimes individual trees, received a sort of headmaster's report (some, such as the maple Rubrum sanguineum in 1930, getting a bit of a ticking off: "It would be hardly worth planting for autumn effect if it could not do better than it has this time"), and the year's display was duly appraised. The question which perpetually fascinated him – and still does at Westonbirt – is the relationship between a year's autumn and the preceding weather. Mitchell's assumption, still shared by many autumn-watchers, is that plenty of summer sunshine – but not to the point of drought – is needed. This, goes the theory, boosts trees' sugar production, and so, when the green chlorophyll dies as days get shorter, the red and gold pigments which have been there all along shine through. It did not take Mitchell long to learn that reality did not always fit such rules of thumb, as these pre-war records show:

1928 After a dry summer with abundant sunshine, autumn colour was "exceptionally good" but short-lived. Week earlier than usual, at its best 7 October, after which the prolonged drought of the summer took its toll, and the leaves "simply fell for lack of water".

1929 (Westonbirt experts think this the year most comparable to 2010.) February snows, and a late, but impressively flowery spring. Warm June, and then an autumn "extraordinarily late... Very little colour was seen until two inches of rain fell in 24 hours on 5 October. In a few days there was a remarkable change". Ended with a gale on 10 November.

1930 Very mild winter, and early spring. Dry early summer, then heavy rains from mid-July onwards. "The finest show was as late as 26 October. A very beautiful and varied lot of colour was in evidence until Saturday 1 November, when a gale the following night and a severe frost on the Monday morning spoilt what had been a fine, although very late, season." '

1931 Mild winter, with cold winds only in late February. Dull, very wet summer. "Until the middle of October, the autumn colouring was very poor." Then, "a succession of white frosts set in, and the effect of this was magical, and from then to the end of November a riot of colour prevailed... an autumn sunset of scarlet and gold with all the intervening shades".

1933 Summer "abnormally dry and hot", so "the outlook looked about as bad as it could do... when, from 7 to 10 October, 1.91 inches of rain fell. This seemed to act like magic, the effect was almost instantaneous, instead of getting a poor season, we had one of the most gorgeous I have ever seen."

1934 Another dry, hot summer, but autumn colour "uneven".

1935 Mild winter, very late, severe frost (11 degrees of it on 20 May), then summer drought and September gale. Average colour, soon ended by persistent wind and torrential rains, save for trees in well-sheltered spots, which lasted well because of absence of frosts.

1936 Wet, occasionally snowy winter, and cold, late spring, followed by wettest July for 50 years, sunny August, rainy September, and dull October. And yet: "There has never been a better all-round colour... I have always thought a wet and sunless summer not conducive to good colouring, but I shall modify this opinion." One maple "dappled with a red akin to a hot coal", another "a delightful salmon flame", and a third "like a lovely piece of tapestry velvet".

1937 Wet, stormy yet mild, with almost no frost. Dull June and July, followed by hot, dry August and September, with no significant rains until late October. "The show kept on and on... Acer Glade was extraordinarily beautiful..."

1938 Very mild winter, and early spring. Summer "one of the most sunless I have ever experienced... the conditions for good colour were all lacking... but it has been a far from disappointing year... this has been the longest season for autumn colour on record."

1939 Bad January snowstorm, and most severe winter since 1928, followed by brilliant spring. "Because of the dull summer we did not anticipate good autumn colouring, but on the whole it has been magnificent... absence of frost was the main cause, and prolonged the show later than I have ever known... Acer saccharum has excelled itself, and for a few days, the best of them were like flaming torches..."

And then came the winter of 1947, the harshest for generations. The damage that the snows and prolonged frosts wreaked on the trees damn near broke Mitchell, who was now in his seventies. His only grandchild, John Earley, son of Mitchell's youngest daughter Enid, reports his mother saying the destruction made Mitchell so distraught he nearly suffered a breakdown. Tellingly, this is the year the log ends. Page 89 of the second of his Westonbirt notebooks is numbered, but never written on. He may have been too busy clearing the wreckage and tending the injured (with a team of men much reduced by war), or too upset to record the damage. Either way, something more than the boughs of his trees seems to have broken that unkindly year.

Mitchell soldiered on until he was nearly 80, never taking a holiday. Lord Morley died in 1951, and, a few years later, his family offered Westonbirt to the nation. The government turned to the Forestry Commission, assuming they would be delighted to possess such a collection. They weren't. Meetings were held, assessments made, cases stated, and yet, when it came to the vote on whether to accept Westonbirt, the commission's board was split. It was only the casting vote of chairman, L Palmer, which saved the arboretum from being broken up, sold, and, almost certainly, cut down. Mitchell was given a small cottage nearby, and lived long enough to see his beloved Westonbirt opened to the public. He was buried in the churchyard of St Catherine's, Westonbirt. His grave has no headstone. It is marked, instead, by a tree. n

For information, visit forestry.gov.uk/westonbirt