At a Saxifrage Society meeting at Wisley last year, I mentioned to Adrian Young that I would be exploring the Lake Garda area in Northern Italy that summer. When I added that I particularly wanted to see Saxifraga arachnoidea, he asked "Why?" in a way that indicated that he doubted my sanity. I had to admit I didn't really know why. By most accounts it seemed a scrawny thing and the Garda mountains certainly have more conventional delights such as Daphne petraea, Physoplexis comosa, Silene elisabethae, and of course, Saxifraga tombeanensis.

In the weeks that led up to my visit, though, it was S. arachnoidea that continued to dog my thoughts. There were three principal contributory factors. Firstly, a full-page photo of it in Herbert Reisigl's Blumenwelt der Alpen that makes it look the equal of any of the most sumptuous European alpines. Admittedly the specimen looks to me as if it has been subtly "massaged" into a tighter clump than would be natural, and the colour values of many of the photos in the book seem enhanced. Nevertheless, it was sufficiently impressive to elevate the species to my desiderata.

Secondly, in The English Rock Garden, Reginald Farrer rather surprisingly chose to eulogise it in the peculiarly rococo style of prose poetry that he normally reserved for his most cherished flowers:

In the damp silty grottoes under the Daphne's cliff on the Cima Tombea this very rare species may be seen, making masses so filmy and evanescent that you expect them to be blown away on a breath like the yellow-starred films of cobweb that they seem. It is as frail in root and stem as the hope of a lost cause, then spreads into a lax flopping jungle of weak fine branches set with broadly oval and minute toothed leaves as flimsy and frail as a lovers' vows; and the whole mass is interwoven with a long soft twining fluff of shaggy silver that gives it the frosted and dew-dropped iridescence of a dream. The little stars of blossom spring here and there on fine stems at the ends of the branches, and are of a pale diaphanous yellow in keeping with the ghostly unreality of the whole apparition.

Lastly, it was apparent from the literature that the species was rare, localised, and elusive even in its known haunts. I had spent 4 days in the region before, not in active pursuit of S. arachnoidea, but checking every likely looking niche I encountered without so much as a sniff. I decided it was clearly a plant a self-respecting plant fiend such as I had to find.

The main expeditionary party consisted of my wife, my one-year-old son, and myself. After the first few days we were joined by my in-laws, who had a very important role to play by freeing me from parental duties often and long enough to satisfy my botanical urges.

Lake Garda is the largest Italian lake and lies at the southern edge of the Alps. It was formed where a giant glacier formerly debouched onto the Po plain. Its altitude is only 65m above sea level, and the climate and immediate flora is Mediterranean. However, limestone mountains to the east and west tightly embrace it, and these plunge dramatically into the water, particularly along the western shore. On the east lies the Monte Baldo range, a narrow ridge with easy access from roads and cable cars. To the west lie a jumble of hills and jagged peaks with more difficult access. These are collectively known as the Tombea range, and form with the ranges immediately to the north and west, the Giudicarian Alps.

Neither range is of any notable height, the highest summit being just 2218m, but they possess a rich alpine flora with a number of interesting endemics. Supposedly this is because, being glacier free, they became a refuge for plant species during the ice ages.

We had flown to Verona, and then were whisked by coach to our hotel in Limone sul Garda on the north-west shore of the lake. Limone is a hotel resort town, popular with Germans, with many hotels and restaurants, and a steamer service across the lake to the foot of Monte Baldo. We found it to be an excellent base for exploring the area. The road beside the west side of the lake is not for the faint-hearted; there are innumerable tunnels, tight corners and a preponderance of oncoming coaches. We had hired a car; this is not cheap, but necessary as a time saver to get to remoter places. It is also considerably less fraught than public transport where you don't speak the language.

Limone is famous for its ruined lemon groves, and these must be a unique example of an economy based upon a pun. The place name is derived from the Latin for a border; this was an old frontier town, and it still marks the boundary between two Italian provinces. However, Limone is also the Italian word for lemon, and, apparently for this reason only, someone 300 years ago thought it would be a good idea to grow lemons here. The climate here is very marginal for this crop, and elaborate walls were built in an only partially successful effort to protect the blossoms from the cold mountain winds. The transport costs to market were expensive too, but until fairly recently there was still apparently a premium to be had for lemons from Limone.

We had chosen the two weeks from 21st June to 4th July, thus coinciding with Wimbledon fortnight. This, we supposed, would attract all the rain in Europe with Centre Court being its epicentre. We were wrong. The first week was a succession of promising mornings, leading quickly to cloud and the onset of steady rain.

On the Wednesday we explored the road described as "rather adventurous" in Webb & Gornall along which S. arachnoidea had been seen in three places in 1985. This road is a narrow, unmetalled track with deep craters, boulders and wash-outs running through a superb limestone wilderness, from the villages that make up the community of Tremosine almost vertically above Limone, up to the Tremalzo pass. It is negotiable in a ordinary car, although you will be in first gear for the most part, and suffer the ignominy of being overtaken by mountain bikes. Furthermore, if your driver is botanically minded, as ours was, their loyalties will be dangerously divided between watching the road surface, and scouring the rocks for plants. One thing is certain; you won't see many other cars!

Suffice it to say that we didn't find S. arachnoidea either. We were particularly convinced by one particular overhanging cliff, dripping wet in places, that we passed underneath as the road curved around a range called Pozza del Lupo. This had curtains of Physoplexis comosa in flower, and the stickily hairy Aquilegia thalictrifolia. We had time enough to pay respects to a colony of Daphne petraea we had seen on our last visit near the Tremalzo Pass before the rain drove us, soaked and chilled from the mountain.

That evening, it was evident that the less botanically inclined party members desired to visit a city, and, for a change, not drive around in the rain all day staring at rocks. Luckily I had a plan. I had photocopied all the pages from Webb & Gornall that could conceivably be relevant to this trip. Among the pages I had included was the entry for Saxifraga berica, although at the time I felt it unlikely to be used. Suddenly, though, it had sprung to the top of my list because a) it grew near a city, Vicenza, and b) this was down in the Po valley, where the weather ought to be less hostile. I suggested that we should visit this historic town, and while we were there, knock off this rare Saxifrage too.

S. berica is a very rare species indeed. It is endemic to a small part of a small range of hills, the Colli Berici. These hills rise 10km south of Vicenza to just 450m, and represent the last hurrah of the Alps before the unalleviated flatness of the Po basin. Even here it is restricted to a rare niche, overhanging shady rocks. The next day, the plan was simple, go to the hills, find the overhang, find S. berica, then have lunch and an afternoon in Vicenza.

We left Limone in a downpour, with the rivers bringing down whole trees that joined rafts of detritus out in the lake. By the time we reached Costozza, the little town beneath the north east end of the Colli Berici, the sun was out, and indeed, it was obvious that here we were in the Mediterranean climate proper. From the town, there are signposts to the scenic track that mounts the hills, exaggeratedly referring to them as monti rather than colli. This track skirts the gaping limestone quarry that spoils the town's view, and ascends in a series of gentle hairpins past imposing villas set amidst tracts of scrubby woodland.

We stopped, and I spent an hour and a half looking for likely overhangs, but found only impenetrable Cotinus and Paliurus scrub, a Scarce Swallowtail butterfly and dry rocky slopes alive with snakes. As a final attempt, my wife suggest I try a track she had spotted that diverged at a tangent from one of the hairpins about halfway up. This track soon doubled back to a house, but a footpath lead on further. Now the vegetation suddenly took on the aspect of a submontane wood; there were Hepaticas, Helleborus niger and ferns underfoot. Then rocks appeared on my left, and a little further a small cliff loomed amongst the laurel trees, no more than 20 feet high and 20 yards long. As I picked my way through brambles to its base, I could see what could only be S. berica here and there along the entire length of the rock.

Looking closer, I could see that my visit could have been better timed; at the end of June, the colony's flowering season had all but finished. There were a few flowers remaining, but the plants themselves were now straggly, and their glandular hairs had caused the tufts to become matted upon themselves. Interestingly, some of the larger specimens were already showing fresh growth in their centre, thus confirming the perennial nature of the species; I somehow expected them to be summer dormant. The overall effect was admittedly not very photogenic, but I imagined that S. berica would put on quite a pretty show of white stars for a visitor in April. One last point I noted was that it grew only on the lower half of the cliff, where there was shelter from the evergreen shrubs, and a sizeable overhang. Even here it picked out only round eroded cavities that appeared slightly moist.

This is not the place for a full description of S. berica; I refer you to Webb & Gornall's book, but the most curious feature is its unequal petals, which it shares only with its closest ally and the non-European section, Irregulares. Unlike the Irregulares, S. berica's petals all differ in size; the largest can be twice the size of the smallest within the same flower.

S. berica's closest relation is S. petraea, a biennial with more deeply lobed leaves and slightly larger flowers. This ranges across the south-east fringe of the Alps from Monte Tombea across to the karst region of Slovenia and Croatia. Its nearest site to the Colli Berici is 60km away on Monte Baldo, and it is not hard to imagine S. berica representing an isolated population that found itself cut off climatically in this interglacial period, and which has adapted by evolutionary means to local conditions.

I did collect a small pinch of the abundant seed, about three capsules full, from S. berica. I passed half on to someone else in the Saxifrage Society, and sowed the rest directly into crevices on a tufa boulder, as a kind of homage to its habitat. I have planted this boulder vertically in gritty compost in a large pot, and kept it in a cool greenhouse, watered from below only. The seeds germinated rapidly; one plant has made rapid progress, and as I write 8 months later, it is opening its first flower. There are about 20 more, much smaller ones on the tufa plus a few evidently blown from their crevices that have germinated in the grit. Those that are on drier, sunnier aspects show signs of stress by a bronze tinge to the leaves, but all now seem to be responding positively to the onset of spring.

S. berica, S. petraea and S. arachnoidea form three-quarters of series Arachnoideae within section Saxifraga, according to Gornall's classification system. Previously, under Engler's system, S. berica and S. petraea were in the Tridactylites section, while S. arachnoidea was a member of the Nephrophyllum group. The fourth member of the series, S. paradoxa, from Styria, was usually placed in the Tridactylites also, but had at one time been granted a genus of its own, Zahlbrucknera.

The 4 species are all endemic to the South Eastern Alps and its environs, and all share the unusual habitat of shady overhanging cliffs, rocks and cave entrances. This niche is not restricted to this group amongst Saxifrages; I have seen S. hederacea and S. chrysosplenifolia in a similar habitat around the entrance to Tzanis cave on the Omalos plateau in Crete. After finding S. berica, my target had now altered from just finding S. arachnoidea alone, to acquainting myself thoroughly with the Arachnoideae.

During the fortnight, we made two assaults on the Monte Baldo massif. We skirted the lake around to Malcesine, and then rode the cable car to the summit ridge. From here, the view is stunning; Garda, 5000 feet below, appears to be directly down beneath your walking boots. The Tombea range across the lake becomes just an undulating plateau, and your eyes are drawn away to the eternal snows of the Adamello range in the distance to the north-west. The view this particular day was, however, short-lived because, in no more than half an hour, the sky had darkened to a leaden hue. Across the cloud base peculiar furrows and striations developed, as the coming storm was whipped up and over the spine of Baldo by the east wind. Lightning ranged nearer and nearer, and we were forced back, half-running, to the sanctuary of the refuge, and retreated back down the mountain before the cable car was shut for safety.

On the first rocky slopes encountered going south, however, I had noted abundant leaves of the Baldo endemic, Callianthemum kerneranum already over, growing with strong colonies of Saxifraga mutata, frustratingly not yet in flower. This was my first ever wild encounter with this weird Ligulatae species whose squinny orange petals it borrows from S. aizoides.

On another day we optimistically packed a picnic, and explored the old military road that runs north-south behind almost the entire length of the ridge along the 1430m contour. On this range in 1911, Farrer collected a plant that is still widely grown by alpine enthusiasts, S. paniculata 'Baldensis'. I had always assumed this was simply the local form from this mountain. S. paniculata does grow in vast abundance on the limestone rocks all along this road, but nowhere did I see any plant that even remotely approximated Farrer's collection, or was anything other than typical. According to its finder Saxifraga paniculata 'Baldensis' grew on the highest summit ridge and so probably represents an extreme microform adapted to the extra exposure.

Where the road passes below Cima Valdritta, the highest point of the chain I noted an unusual form of S. aizoides, dwarfed and growing as a limestone crevice plant, and with orange flowers. The star plant at this point though was undoubtedly Paeonia officinalis. We had seen the leaves in several places around the mountain, but here I could at last scramble the 10 feet up the roadside rocks, gaze into its rosy chalices, and drink their sweet scent.

On the road down, we passed beside a small reservoir at about 1000m by the village of San Valentino. Just below the dam, in a gorge, my wife suddenly told me to stop because she had spotted a flower she didn't recognize on the wet rocks. A swift piece of reversing proved itself worthwhile when the mystery plant was recognized as S. berica's biennial sister, S. petraea. Here there was a small colony of some 25 individuals in full flower scattered beneath an overhang around a culvert. According to Webb & Gornall, as well as Winton Harding, this can be a very worthwhile plant in cultivation, but Farrer rather unnecessarily noted it was rare in nature, but even rarer in gardens, and deservedly so. With its softly hairy leaves recalling a jagged Dovesfoot Cranesbill and slightly unequal notched petals of pure white, I deemed it dainty and wholly charming.

With the available days fast dwindling it now became imperative once and for all to locate S. arachnoidea. This would demand a full day's botanical exploration in its heartland, the Tombea range. My in-laws are game for most things; I once persuaded them on the first day in the mountains to climb 4000 feet to see S. florulenta in the Maritime Alps, by selling to them the concept of The Ancient King as being the plant equivalent of the Giant Panda. The Cobweb Saxifrage was no Giant Panda though; its notorious elusiveness and the modest reward its discovery confers suggested a kinship with the quarry of The Hunting of the Snark. In the end, a day spent with their grandson on the steamer to Malcesine proved a greater attraction, and I was left to my own devices for a day.

Some references deny that there is any such mountain as Monte Tombea. This is not entirely factual. The highest point in the Tombea range is Monte Caplone at 1976m, set in its dead centre. This is the highest summit of an east-west ridge shown on old maps as Monte Tombea; current maps name a subsidiary top Cime Tombea. On its southern flank the headwaters of the Fiume Toscolano gather, and this drains the Val Vestino, the largest upland valley between the lakes Garda and Idro. Webb & Gornall state that the upper part of the Val Vestino is the region in which S. arachnoidea is commonest, so this was to be my target.

The drive there from Limone is a scenic one taking an hour and a half at best. I headed down the lake to Gargnano where a side road snakes steeply up the wall of the Garda basin and emerges in the valley above. I followed the valley bottom road past the large reservoir that now occupies its middle section, and signs were taken towards Magasa, a hilltop village where the roads stop.

On the approach to Magasa the route hairpins across a tributary valley, and passes beneath an overhanging cliff festooned with Physoplexis comosa. A perfect S. arachnoidea niche, I thought; for such a local plant S. arachnoidea has a large altitudinal range, colonies being known from near the summits right down to 600m. The dry silty underhangs in this case, though, proved to be exclusively occupied by Silene saxifraga and Moehringia bavarica var. insubrica. This latter is a sandwort with glaucous leaves and belongs to a genus that, like the Series Arachnoideae, has a number of local species in the southern Alps that specialise in this unusual habitat.

Just before Magasa a side road with hairpins so tight they required 3-point-turns, heads up to a community called Denai, and here I parked with Monte Tombea before me like a sleeping giant. I walked passed fields infested with Hellebores towards the mountain, then followed path 67 eastwards through dense forestry. On and near rocks in this wood I again found S. petraea, this time a stronger colony with outliers 300m away. Most plants were in overhung rock crevices, but others had thrived in deep detritus amongst other vegetation, under the shadows of the trees. Still others grew on small boulders in an adjacent field, with only a vestige of overhead protection. This species was clearly not as fastidious in its choice of habitat as the others of its section.

My route then almost doubled back onto path 66, which heads 600m up to the Tombea ridge. This is an excellent walk. It first leads under a large bluff again adorned by Physoplexis comosa, but this time accompanied by large mats of Daphne petraea, which for the most part, appeared not to have flowered. Half way up a stream is crossed, on whose banks grow the unusual Leuzea rhaponticoides, resembling a giant Knapweed. The woods give way to turf with Primula spectabilis, now finished, and Viola dubyana where the soil had been disturbed.

Just as things were getting interesting, the clouds descended, rain and hail began, and thunder was heard. I made for the gorge of the stream, declared myself an early lunch, and took shelter, like S. arachnoidea itself, beneath a rocky overhang. An hour later I had had time to consider why S. arachnoidea had adopted such a peculiar habitat. Any evolutionist would no doubt find this question itself wrong-headed: S. arachnoidea didn't adopt the habitat, it is nearer the truth to think of species as "children" of their environment. They are created and moulded into winning designs over generations by the selection pressures of the ecological niches themselves. However, as I sat there huddled against the wind and rain so intense that the meadow herbs appeared to be both waving and drowning, it occurred to me that the title I have given this article was possibly a more satisfactory answer.

For once, the rain eventually relented, and I continued my climb, shortly reaching a T-junction of paths. This is the Sentiere Antonioli, an old military road, now overgrown, that follows a contour below the Tombea ridge, and then continues over hill and valley to the Tremalzo pass. I headed east along it, a path marked number 69, towards the Bocca di Campei, the col between Cima Tombea and Monte Caplone. A hundred yards or so along this path my eye was caught by a colony of the common Saxifraga caesia, and what appeared to be a giant flowered form in its midst. An instant later I realised that I was looking at S. tombeanensis, albeit a small specimen with only 3 flowers.

This classic Kabschia species with green spiny leaves and small cymes of pure white is of restricted distribution, but extends beyond the range for which it was named north into the Brenta Group. Another Kabschia, S. vandellii, is also known from the Tombea massif, but I did not see this. I worshipped my find briefly, then sought a better specimen. This search proved initially fruitless apart from finding a solitary plant of Silene elisabethae, another one of the superior endemics of the region, with its giant ragged-robins sitting atop glossy, sticky leaves, and Saxifraga mutata, as before not yet quite in bloom.

The path sweeps around north at this point, and heads towards an archway blasted through the rocks. Before reaching this, however, I passed a slab of limestone on my left. I could see it overhung its hidden base, so I clambered up to it. As my eyes came level with its foot, I finally came face to face with S. arachnoidea. Set back four feet under the ledge were about 8 plants in their prime, growing in pure lime dust eroded from the slab, and sharing this specific habitat with no other species. In less shady parts of the overhang though were the Silene saxifraga and Moehringia bavarica var. insubrica I had seen earlier down below in the valley.

S. arachnoidea's pretensions to the higher echelons of alpine society were reinforced by the aristocratic nature of its other closest companions: within 5 feet grew both Physoplexis comosa and Daphne petraea, and the rosy flowers of the latter could be seen 20 feet straight up. And, 8 feet above, a rocky protuberance was the throne for a huge S. tombeanensis fully 2 feet across, covered with developing fruits. For me, though, as you will have guessed by now, there was only one monarch here, and I prostrated myself before the Cobweb Saxifrage, and took enough photographs to satisfy a member of the paparazzi. Then I examined with wonder the minute details of the bizarre leaves wrapped in cotton-wool through my hand lens.

The world distribution of S. arachnoidea is tiny; all the colonies are within 18km of the nearby village of Storo. I have demonstrated to my own satisfaction too, that it is not at all common within this area. There is a suggestion in the literature that the colonies are dynamic, and are not to be found in the same places for long. If true, this is surprising; its habitats are fairly static and stable, and free of competition. It is also hard to imagine a mechanism for conveying seed from one suitable ledge across kilometres of unsuitable land to another ledge with any degree of efficiency. Perhaps instead, seed reservoirs are present in the substrate in the known sites, and only germinate in years when the microclimate suits. The individual plants are perennial but short-lived, so colonies would seem to die out and appear elsewhere.

One detail I noted differs from Webb & Gornall's description; they state the flower colour is off-white whereas other sources say yellow. To my eyes, and judging by my pictures, they were clearly yellowish tending if anything towards green.

I would have to confess that I was disappointed with the quality of my photos of the Arachnoideae. I had expected to be taking pictures of alpines in the sunny south Alps, and ended up photographing plants from extremely shady habitats in the rain. I routinely use 100 ASA film for its finer grain, but in these circumstances it is just too slow. Anyone following in my footsteps is advised to take a second camera body, if possible, loaded with 400 ASA film. I certainly intend to do so when I eventually visit the Koralpe range in Austria, for visit it I must in order to find Saxifraga paradoxa, the remaining and dingiest Arachnoid Saxifrage. Perhaps, Adrian, you were right after all to doubt my sanity!