QBARS - v20n2 Two Tourists in Japan: Part II

Two Tourists in Japan
Frank Doleshy, Seattle, Wash.

Part II: Rhododendrons kiusianum and yakushimanum

Part I of this series was sent to the Editor with a few loose ends. This was unfortunate but hard to avoid; we had collected seed with the idea of distribution by the American Rhododendron Society Seed Exchange, and the urgent matter was to publish some notes before the distribution date.
These notes raised a question about germination of the not-quite-ripe seed. However, for us, this has actually grown as well or better than the ripest. Also, no definite name was given for the species collected as No. 13 and No. 15, because the various authorities did not agree. To clear up this matter, Dr. Sleumer, of the Rijksherbarium, suggested that we obtain a paper by Tor Nitzelius, of the Botanic Garden, Goteborg, Sweden: "Notes on Some Japanese Species of the Genus Rhododendron," in Acta Horti Gotoburgen sis, Vol. XXIVA (1961), pages 135-174.*

*Also translated to German and published in Rhododendron Jahrbuch 1963, of the Rhododendron-Gesellschaft, Bremen, Germany.

Mr. Nitzelius kindly furnished a copy of this important study (in English). Here, he convincingly argues that R. brachycarpum and R. fauriei (or fauriae ) are two names for only one species and that R. brachycarpum is the valid name. Therefore, we suggest the use of this name for No. 13 and No. 15. Actually, there is no conflict here with popular British and American usage: both of our collections came from plants with indumentum, and this is commonly thought to be the distinguishing feature of R. brachycarpum .
Mr. Nitzelius also reviews the other Japanese Rhododendrons in the Ponticum Series, and his conclusions as to classification and naming are extremely interesting. However, many of the people who read this Bulletin are amateur growers who perhaps have some knowledge of the Japanese Rhododendrons and would like to find out more about them without first having to learn a new set of names. Therefore, with apology to the more professional readers, we delay the introduction of these new names until the final article appears.

The Journey Resumed
Returning now to the south island of Kyushu, we enjoyed a visit to the Aso Volcano but did not feel qualified to make any accurate seed collections from the various azaleas in that area. Then, rolling down the Pacific coast on a semi-express train (by now, a mode of transportation as comfortable as an old shoe) we felt a twinge of conscience when we watched the only other foreign couple get off at Nobeoka. We knew of an alluring rhododendron population in the mountains above but had not been able to schedule time for a visit.
Our destination was Miyazaki, which had merely been an unfamiliar name three months earlier. But our travel agent had said, "You might like the place. That's where it all started Japan, I mean. Of course nobody speaks English, and you won't see any other Americans." Thus we were led away from rhododendrons and into the midst of Betel Palms, Phoenix Palms and fantastic shorelines, but the end result was a surprising encounter with a rhododendron.
Arriving at Miyazaki, we found that it was not only a center of archaeology and legend but also a friendly, modern place. After lunch in our room, we decided to see if the inn's staff could tell us anything about rhododendrons. Our Japanese vocabulary was still much limited to "Shakunage," the word for rhododendron, and verbal communication was not successful. But the manager could read our notes and write back. He was somewhat nonplussed at tourists who asked about Shakunage while in banana country - and, from then on, all the maids merrily greeted us as "Shakunage-san." However, he got out a travel folder for the Kirishima Mountains, written entirely in Japanese. This contained an interesting photograph. Taken in dim, blue light, it showed a mountain plateau with many compact shrubs, bearing great quantities of pink and rose flowers. We could not make out the individual leaves or determine the scale, but we were sure that this was a rhododendron. "Shakunage?" we asked, but the manager wouldn't quite agree. "Perhaps close to Shakunage. You should go see it." Nothing could have kept us away, and we obtained the exact location: Ebino Plateau, a favorite mountain retreat of the Satsuma lords, located at an elevation of 1200 meters in the west part of the Kirishima Mountains.
Three days later, via bus from the city of Kagoshima, we left for Ebino. This bus went only part way, and we had a 20-minute stop - over at a mountain toll gate. No other people were waiting there, and the toll collectors apparently thought we looked lonesome; they invited us to come in and use their office as a lounge. But there was too much to see out in the fall sunshine. Big red-barked pines (Akamatsu) were lightly decorated with the red leaves of climbing vines, producing an effect not at all like the massive ivy sometimes planted around trees in this country. Beneath, there were evergreen shrubs among the fall foliage, but nothing we could recognize as a Rhododendron. Not high enough, we decided, and got aboard the uphill bus. But our score was still zero when we arrived at Ebino, so we climbed further, among small pines which are apparently sheared clean of dead limbs by the winter ice storms.
Finally, creeping on the ground, there was an indisputable rhododendron of the evergreen azalea group, but we could find none of the big game (such as R. metternichii ) which we still thought we were hunting. Truth finally dawned when we went back down to the Ebino pumice flats; the little creeping rhododendron seen among the pines was growing in the form of dense shrubs, to knee-height or a little higher. This was the plant in the photograph, and we found the place where that had been taken. Thus we met R. kiusianum , the Miyama Kirishima, which attracts tens of thousands of visitors to these mountains in May.
We had never been able to see anything very interesting about the hundreds of evergreen Azalea hybrids which flood the sales lists. But, face to face with genuine R. kiusianum , we were shocked to think that we might have ignored it. This compact little plant with purple-bronze fall foliage is possibly the prize of our entire trip, and it taught us a lesson about trying to judge the quality of a plant by its hybrids. The seed was ready to shake into the bags without even picking the capsules. So we obtained a good supply as our No. 8, and have since been delighted at near - 100% germination.
We do not recall any commercial offering of this plant, and it seems practically unknown to Pacific Northwest gardeners. In The Species of Rhododendron it was considered part of the R. obtusum complex, but it has since been given individual status in The Rhododendron Handbook . It is listed without comment or rating in Rhododendrons for Your Garden . Therefore we give the following details: On the open Ebino pumice flats some plants are rounded and others flattened out to a shape more like a drum or a cake. Few if any are over 70 cm. tall. And, in the woodland, they are usually lower, forming a mat a meter or more in diameter and perhaps one-fifth as high. All are very bushy, with terminal twigs 1½-3 cm. long. Such a twig carries 4-12 leaves, somewhat clustered toward the tip. Leaves are 4-11 millimeters long by 2-6 millimeters wide, varying from elliptic to obovate, often narrow at the base and spoon-shaped at the tip, and always ending with a small, blunt mucro. Under a lens, the top surface appears shaggy with long, pale hairs, but the undersurface has few hairs except along the midrib. Twigs and buds are more hairy than the leaves. The hairs are "footed" with small platforms of hard tissue which look deceptively like scales where the hairs are broken off. In October, only the exposed leaves were bronzed; those in the woodland remained bright, deep green, lighter beneath.
We did not see flowers, which are said to range from pink through rose to purple, with a rare white or scarlet specimen. Photographs indicate that there is often a flat or gently curved sheet of flowers across the top of the plant.
Our dimensions are smaller than those given in the 1963 Rhododendron Handbook or those implied in The Species of Rhododendron . Perhaps we were in the upper part of this plant's altitude range, and it will be interesting to see if the small size persists in cultivation. For a companion plant, we can make a good recommendation which might he difficult to follow: Blue autumn flowers of Gentiana triflora were growing up through the R. kiusianum on the plateau, with an effect which would make any gardener happy.

Off To Yakushima
Two mornings later, at Kagoshima airport, we ordered coffee in the sunny lounge and wondered what kind of a bird a 16-passenger De Haviland might be. When Toa Airways No. 531 wheeled up to the gate, we were surprised to see four big engines-reminding us that this is typhoon country. The stewardess didn't really expect us to be able to read one of her morning newspapers but wanted us to see the beautiful capes and islands below. After each loudspeaker announcement in Japanese there was an English translation, followed by a smiling visit to our seats. After 35 minutes we descended over surf and sugarcane to Yakushima airport and got a taxi for the half-hour ride to the Yakushima Kanko Hotel, at Anbo. As we had learned before, the "hotel" part of the name didn't really mean that we had to go back to knives and forks and all that business. We were in a magnificent Japanese inn, built only three years earlier. But there was no indication that they had received our letter about going to the mountains. So, to introduce the topic, I put on outdoor clothing, slung the camera over my shoulder, and went to the entrance hall.
At first, the language problem seemed dense. However, the junior high school social science teacher arrived by taxi, with another teacher in tow, and introduced himself in good English. We found that they weren't exactly expecting Shakunage enthusiasts. But the teachers and inn staff settled down to discussion and produced sketches of one of the three mountains, Kuromi-dake, with every possible arrival and departure time noted at summit and base. After two hours and a change in air reservations, we had a plan, and went to our room with the teachers for lunch. Dietary conservatives may find it comforting that the teachers didn't care much for the sashimi course, from a large, live lobster in the middle of the table. But everyone enjoyed the boiled lobster.
We learned that the Yaku residents emphatically do not regard their beloved high-altitude Shakunage as something to be weeded out of tree farms. They were aware that it was grown in England and proud and happy to learn of continued interest in North America. Dr. Serbin's 1959 visit was well remembered, not only because of his successful climb to the Shakunage country but also because of his friendly encounters with the school children.*

* For accounts of Dr. Serbin's visit, see his articles in this Bulletin, January, 1960 and April, 1965. Also of interest are extensive references to Dr. Serbin's observations in the Nitzelius paper, cited above.

After lunch, we found that Mr. Tsumagari - the gentleman pictured on the cover of the January Bulletin had decided to go with me. Also, another member of the inn staff volunteered to come and carry our supplies. I was glad to have companions, because this was the one place in Japan where Mrs. Doleshy and I divided forces. The 4:00 A.M. departure from some unknown point in the mountains sounded too much like the drag up Mt. Rainier, and she was staying at the Yakushima Kanko.
Everything was ready, so we got in a taxi and headed for the narrow gauge logging railroad made famous by Dr. Serbin. This is not exactly like an air-conditioned 2 kilometer-per minute Bullet Train on the New Tokaido Line, but we seated ourselves on packing cases in what might be called the observation car and started uphill at 3:30 P.M. (See Fig. 20). The trip might as well be enjoyed, and, if one gets tired of looking for the bottom of the granite abyss on one side of the track, there are lots of tree ferns on the other side. After about 45 minutes through bewildering overlaps of tropical and temperate flora, we got off at the main uphill station, Kosugi-dani. Here, we walked into a comfortable inn, but English communication was nil, either verbally or in writing, and I wasn't even sure we were staying there for the night until I gathered that it was time for me to accept the honor of first bath. Then, when we were starting dinner, a cordial gentleman walked in from the next room and introduced himself in English as Professor Dr. Sohma of the Biological Institute of Tohoku University-formerly a student at the University of Arizona and the University of Oregon, now deputy director of the Tohoku University Arboretum and an authority on fossil pollens. He was on his way down from Kuromi-dake and wondered if he could perhaps help with my specific interests. Conversation started with Rhododendrons, reached Salix during the lobster course, then branched off in all directions. This enjoyable evening ended at 9:00 P.M. because we planned to be on the trail at 4:00 A.M. - and were, traveling by flashlight for the first two hours.

Passengers on the Yaku train Yaku Train
Fig. 20.  Passengers on the Yaku train.
Mr. Tsumagari at left, author at right.
Fig. 23.  Rolling downhill on our special train.

Up The Mountain
We followed or short-cutted railroad tracks for several kilometers then headed into steeper country. Somewhere on the way, our route diverged from Dr. Serbin's route to Hananoego, but the two trails are doubtless similar. Ours was like many of the mountain routes in Washington and Oregon. Impassable for any kind of pack animal or wheeled vehicle, and requiring caution on foot logs and wet rocks. However, such routes are not so common in the eastern U.S., and most readers from that area would find the trail to be exactly as described by Dr. Serbin - somewhere between miserable and diabolical.
The route led between Cryptomerias with trunks up to 3 meters in diameter and tops stubbed off by typhoons, forests of Pieris, many Azaleas, and hundreds of things unrecognized. The small, scale-like leaves of Cryptomeria were often in such long strings that the trees looked like giant pines. Shakunage appeared at 1500 meters, as Dr. Serbin found. The first plant was a magnified version of the popular image of this species; leaves were widely-spaced, 15-20 cm. long, and quite flat. But other plants in the immediate vicinity, with more exposure, had the familiar small, convex leaves. Also, a single plant often carried a wide variety of leaves, depending on the exposure of the individual branches. Both here and above, indumentum color varied from pale buff to a rich orange-buff.
Slightly higher (ca. 1550 m.) the trail zig-zagged up rocky terraces, 1015 m. high, with very ample water on the flat steps, and we found huge plants with a slanting height of perhaps 4 m. and trunk diameters up to 28 cm. These were never very upright, and the big, bare stems sprawled out above the rocks and the trail. Plants of this size were not seen by Dr. Serbin on the Hanano-ego Route, and they probably are found only in particular types of terrain. On the other hand, I did not see any plants of the form which has nothing but small, nearly-circular leaves. Individual leaves of similar shape are common enough on cultivated plants of R. yakushimanum and are probably the ones which emerge very late in the season.*

*Mr. B. M. Swenson of Seattle has suggested this reason for the round-leaf abnormality often seen on members of the Thomsonii and Falconeri Series, as well as on R. yakushimanum .

However, Dr. Serbin, in his articles and correspondence, tells about a group of plants bearing this type of leaf only. They were on a small, wet, mossy plateau which has since been washed out, and it is to be hoped that this rare variant can be found elsewhere. Also, I did not see any leaves without indumentum, a variant found by Dr. Serbin in plants grown from seed collected in the wild.
As mentioned by Dr. Serbin, the plants in native habitat do not retain their leaves as long as those in cultivation. Instead of 7-9 years, the average is probably about 4. I think that this is at least partly explained by the typhoons; a leaf more than 4 to 5 years old probably doesn't have a chance. At higher elevations the leaves are also, in most cases, close together at branch tips, on very short annual growths.
Going on up the trail, we continued to encounter Shakunage, with interruptions only in the flat areas where it seems unable to compete with Pieris and other plants. Average leaf size decreased gradually with altitude, but there was no distinct break. The environment of the individual plant appeared to be the main factor, and shaded leaves at 1650 m. were generally larger than exposed leaves at 1500 m. (See measurements, below.)
I was interested in getting seed from low-elevation as well as high-elevation plants. However, below ca. 1600 m., I found not a single capsule, and there were only a few on the plants from 1600-1700 m. (collected as No. 9). This is in line with Dr. Serbin's observation that the lower plants are not the most floriferous. However, it should be recognized that these observations do not conflict in any way with Mr. Wada's impression that the lower forms have more flowers per truss and are therefore more showy.*

*'See Editor's notes in Oct. 1964 issue of this Bulletin, pp. 197-198. Also confirmed in correspondence from Mr. Wada.

At the time, I suspected that some of the lower plants may originate from seed blown or washed down from the better seed-producers near the summit. This may or may not be correct. However, in recent correspondence, Dr. Serbin indicates that seed and seed capsules are indeed removed by the strong winds. In September, 1965, one of his correspondents visited Miyanouradake (2 kilometers from Kuromi-dake and 99 m. higher), and found ample seed. But, a month later, no capsules could be found at the same location.
The top hundred meters of Kuromidake is in the domain of the full-force typhoon, aiming for destruction at Kyoto, Nara and the Plain of Kanto. All vegetation in this summit area looks sheared off or pressed down, and we found the great thickets of dwarfed Shakunage, with an average height of 2/3 to 1 m. These often grow beside and through conifers of the same height, but they also form solid clumps 3 m. or more across. My photographs also show beautiful red and orange foliage on companion plants. This amazes me; while on the spot and trying to make multiple use of every minute, I didn't see these other plants, or at least can't remember them. Also, I missed the dwarf R. keiskei , a form which Mr. Wada has found very desirable because it keeps its low stature and is not bothered by the summer heat in Tokyo.
We climbed the large double block of rock which forms the summit and, in a pocket of soil, found the highest Kuromi-dake Shakunage. Then, on the shady, sheltered side of the summit, I descended about 10 m. to a ledge and was startled to see one or more large specimens, possibly 2½ m. tall, apparently much like the plants down at 1500-1600 m. Unfortunately I did not stop to obtain photographs or herbarium material-an omission which I regretted when trying to piece together the factors affecting leaf size.
On the summit ridge and everywhere else, leaf color was an exceptional deep, bright, glossy green. Capsules were covered with silvery or pale buff tomentum, like that on young leaves. On picked capsules this turned brown-accounting for the brown capsule color noted in descriptions. However, the pale color of the fresh capsules blended with the color of the buds and the light soil and acted as camouflage. The capsules were not only hard to see but also scarce- completely lacking on 19 out of 20 plants in the summit area and even less abundant at lower elevations. But, with Mr. Tsumagari's help, I obtained enough of the high-altitude No. 10 to send Mrs. Berry a respectable package for the Seed Exchange.
The soil all the way from Anbo to the summit is derived from decay of a granite-type rock. Near the summit, this soil is often a thin layer, composed mainly of rock granules. And, because of this scarcity of finely-divided material, the acidity is probably affected by such local conditions as the amount of humus and the relative dampness. Dr. Serbin mentions a pH of 6.2, presumably for Hanano-ego soil. But my very granular sample, from a Shakunage-covered slope just below the summit of Kuromi-dake, gives a reading of 5.3-5.4.
On our clear October day, a pond at ca. 1600 m. was covered with ½ cm. of ice at 8:00 A.M. Other features of the climate are explained by Dr. Serbin, and, since he notes that rain is to be expected 35 days per month, I can only say that I was very fortunate to climb on October 36th. More seriously, I know very well that he gives a correct picture of the weather, but I happened to strike one of the rare autumn days with 100-kilometer visibility. People seem interested in comparing this habitat with something more familiar, and the best analogue which I can suggest is an extremely wind-swept, thin soiled hilltop in the panhandle area of SE Alaska. Also, there is a vague resemblance to parts of the Washington and Oregon coast, but this is quite a far-fetched comparison.
Shakunage seems to be limited to the portion of Yakushima above 1500 m. elevation, and the total area of this high country is only about 16.29 square kilometers, or 6.3 square miles-precariously small in comparison with the thousands of square miles where R. macrophyllum is found. However, at least on Kuromi-dake, there was no evidence that plants had been stripped off by collectors and very little evidence of any other abusive treatment. The typhoons are a natural enemy in that they remove pieces of the plants. But, on the other hand, they probably ensure survival by keeping the habitat open, limiting the competition of Pieris and other plants, and distributing the seed to all possible growing sites. In fact, it may be accurate to say that the Shakunage population survives because of the typhoons, and expands or shrinks with long-term changes in the average velocity and track of the winds. The plants appeared very healthy at all elevations, with only minor insect damage and an occasional white-spotted leaf.

R. yakusimanum R. yakushimanum
Fig. 21. R. yakushimanum at foot of summit
rocks.  Muromi-dake.  Dense layer
of leaves faces the sun and the
Fig. 22. R. yakushimanum thicket photograph
from lee side, showing leaves clustered
at stem tips. Summit ridge of

Individual Variation
A much-discussed question about R. yakushimanum is the amount of natural variation. That is, are the differences in hereditary make-up important enough to justify the naming of distinct varieties or forms?
This is partly a matter of defining the terms. According to one widely accepted view (which I share), a "variety" is a segregated and self-perpetuating plant population which usually does not interbreed with other varieties because of some geographical barrier or some difference between habitats.*

* Concept; of the species, the variety and the forma are discussed with clarity by Arthur Cronquist in the introduction to Part 5 of Vascular Plants of the Pacific Northwest, University of Washington Press, Seattle, 1955.

On this basis, the thousands of plants which I saw couldn't very well be split up into varieties. They were not growing in segregated groups, and it was easy to find intermediate plants which bridged the gap between the more distinctive specimens.
Major variants within a continuous population may, however, be distinguished as forms (or forma). Dr. Serbin's plants without indumentum or his round-leafed plants would be recognizable forms, as much entitled to separate identity as, say, the plants of R. japonicum which have leaves with white undersurface.
I found nothing unusual enough to suggest the establishment of a separate form. However, as shown in the following summary, the leaves are far from uniform:

Group of Specimens:
No. 10 - Leaves from plants in the 1720-1836 m. altitude range which actually supplied the seed so numbered. These were generally the exposed, small-leafed plants; more sheltered plants had larger leaves but no seed.
No. 9 - Leaves from plants in the 1600-1720 m. range which actually supplied seed.
Upper trail - Sample from about the same altitude range as No. 9, but taken from plants encountered at random. Includes leaves from more sheltered plants, without seed.
Lower Trail - Same, but from ca. 1550 m. to ca. 1600 m.
Leaf Data: No. 10 No. 9
Number of leaves collected. 26 3 6 8
Length of blade, in cm. 3.0 - 7.8 5.8 - 9.9 5.4 - 12.3 6.0 - 12.7
Width of blade, in cm. 0.9 - 2.8 2.0 - 3.4 2.1 - 4.1 1.7 - 3.8
Number of very broad leaves, length about 2X width, usually widest close to tip 5 of the ten leaves on one branchlet; 1 of the six leaves on another. None None None
Number of very narrow leaves, length at least 3X width. 11 None 3 6
Number of intermediate leaves. 9 3 3 2
Number of tightly rolled or
very domed leaves.
(more rolled than domed)
2 3 7
(more domed than rolled)
Indumentum variation (from the most common type, which was thick and loose-surfaced, in grayed hues of Spanish Orange, HCC 010):
No. 10-Light buff on 5 leaves; dark brownish on 1; largely worn off of 1, apparently by rubbing on another plant.
No. 9-Distinctly spongy on all 3 leaves; light-colored surface, but much more deeply colored in the many small pits.
Upper trail-Very smooth surface in 2 leaves; spongy on 2; light buff color on 1.
Lower Trail-Somewhat more yellowish (HCC 09 range); very pale buff on 2 leaves; distinctly gray on 1 (possibly an old leaf).

These variations are the result of interaction between growing conditions and inherited characteristics, and the relative importance of these two factors requires further study. I can only state my impressions in general terms, as follows.
Plants growing side by side under the same conditions were usually similar, with only minor indications of individual genetic differences. However, the very flexible response to local environment was always apparent (particularly in the case of the large plant on the sheltered side of the summit rock), and this seemed to account for most of the leaf differences. A Seattle incident tends to confirm this view: Plants of R. yakushimanum were moved across the city from an exposed situation in B. M. Swenson's garden to a sheltered location in K. T. Lauhon's garden, and the result, after two years, was noticeably larger, flatter, more widely-spaced leaves-much as if the plants had been moved from 1750 m. to 1600 m. on Kuromi-dake.
Comparing mountain-top plants with those down at 1500 m., I could not see any differences which would indicate that genetic characteristics are sorted by altitude. However, some degree of sorting seems possible, because of harsh weather above and difficult competition below. Also, Mr. Wada advises that plants from mountain-top and woodland sources are morphologically different when cultivated at his nursery in Yokohama. We should know more about this when seed from different elevations has been grown by many observers. This was the reason why I obtained No. 9 and No. 10 as separate collections and split the small amount of No. 9 into so many shares. Also, available data may be obtained by another Seattle member of the American Rhododendron Society, who plans to be on Yakushima about when this Bulletin is published. He has the advantage of being able to speak Japanese, and he may use Kosugi-dani as a base for exploration of all the peaks.
The equally interesting question of this rhododendron's standing as a separate species will be taken up in Part III; we now return to civilization.

Down The Mountain
Leaving the summit, it was obvious that we couldn't get to Kosugi-dani in time for the regular downhill train to Anbo, and I was wondering how I was going to catch the 3:45 airplane to Kagoshima. Mr. Tsumagari smiled and urged good cheer when I pointed to that time on my watch, but we walked quite fast. Finally arriving at the upper outskirts of Kosugi-dani, we borrowed a 6-man wooden car, complete with hand brake, from the woodshed of a hospitable young couple, lifted it onto the track, and rolled to the main uphill station. There, after several phone calls, we were designated as a special train - Mr. Tsumagari piloting. (He knows everybody!) We rolled for about 5 minutes, then had to wait at a switch for a scheduled uphill logging train. The sun was out, we were well stocked with refreshments, and the picnic would have been great if it hadn't been for the airline schedule. The other train eventually came, and we went down in a magnificent swoop, powered by gravity (very ample here) until we were about 1 kilometer out of Anbo. Another train was in full possession of the track, loading logs. After about 10 minutes and some discussion, they backed off, and we rolled into Anbo, jumped into the waiting taxi, and reached the inn 50 minutes before plane time.
Mrs. Doleshy had given up hope of catching the 3:45 and wasn't quite packed. She had been taking care of such duties as being interviewed and photographed by the Yaku press, drinking Coca Cola furnished by the Yaku bus company, and meeting a Yaku collie by the name of Lassie, owned by a little boy who wanted to show her a butterfly. Also, she had looked for seed capsules on an unidentified azalea growing around the inn, and this became our No. 11
The relaxed atmosphere changed quickly. I had to wash the granite dust off my face, change clothes and pack, while she convinced the inn management that foreign guests should indeed pay their bill just like anyone else. The maid was trying to give me a cup of coffee and help me change clothes-and didn't even have time to point to the beautiful lunch on the table - while two assistant managers were trying to drag me out bodily to the waiting taxi. The driver got us to the airport with at least 5 minutes to spare, and we ate late lunch or early dinner at Kagoshima, caught our train, and slept in Kumamoto that night. Sakae Tominaga, assistant manager of the Hotel Castle, had stayed on duty very late to make sure that her foreign guests were treated like feudal lords, and I think that an unshaved Shakunage-hunter didn't quite fit the image. But, after two days in this likeable town, we looked a little better and were ready to go far north to the home of R. degronianum and R. brachycarpum .