SULCOREBUTIA, FOOD FOR TAXONOMISTS ? (Part 2)

Johan Pot

Fortunately Nol Bredero and John Donald Günter Hentzschel (1999) also emended (1981) were able to make the final pro- the genus Sulcorebutia. He thought the nouncement. They wrote enthusiastically : shape of the scales on the pericarp to be of „Ein Problem ist gelöst!“ Brederoo had major concern. Only the scales of the ge- discovered that in case of sulcorebutias, nera Weingartia and Gymnocalycium were tiny hairs are found behind the scales on similar to those of Sulcorebutia. A second the pericarp, which never occur in wein- characteristic was the non-branching fu- gartias. The consequence was a recombi- nicles of Sulcorebutia as opposed to the ning of Weingartia purpurea and toro- plural branching of the two other genera. torensis to Sulcorebutia. The sulco- This suggested a close relationship bet- collectors were satisfied because now it ween Weingartia and Gymnocalycium. had been concluded unambiguously, that Alas the observation of the plural bran- Brandt was wrong. I was so naive as to ching of funicles in fruits of weingartias check this characteristic and offer my turned out to be an illusion, as was later findings to a group of specialists. I nar- confirmed by Hentzschel. rowly escaped excommunication but they Consensus classification advised me to visit a reliable optician ur- Donald worked on his concept of clines. gently. The characteristic was cherished That’s why he took all plants in the wider for years in circles of sulco-friends. But surroundings of Cochabamba to belong to these days even the most fervent adherents S. steinbachii. It was also part of the spirit will have recognized that most of the sul- of the times to reduce the number of cos don’t have hairs behind the scales on names. In 1984 Donald participated with the pericarp, while some weingartias, for the Huntington Botanical Expedition Boli- example HS 164, do.

Fig. 7: Sulcorebutia steinbachii JK 94 Fig. 8. Flowersection Sulcorebutia steinbachii G123 earlier work ? The editors of Succulenta seem to have sup- ported this opinion for a while. With unbelieving astonish- ment circles of amateurs heard that the members of the work- group looked for consensus and found it by the raising of hands. Was this science or some pragmatic acting in which a majority prescribed what names everyone else was allowed to use ? Would some- body dare to doubt the exper- tise of the members of the work group ? The latter had Fig. 9. Cladogram according to Buxbaum indeed all won their spurs. It was said about Backeberg that he recogni- via. In his list of fieldnumbers of 1989 still zed a not-yet described plant in an unfami- seven varieties of S. steinbachii were men- liar collection immediately. Would all tioned, namely steinbachii, glomerispina, members of the workgroup have such tuberculato-chrysantha, gracilior, kimna- knowledge at their disposal? Or did they chii (n.n.), horrida and polymorpha. have other skills by which they could According to the Cites Cactaceae Check- create an acceptable project together ? list of Hunt (1999) all these names of va- This question had been answered before, rieties should be allowed to expire and in for example by Reto Nyffeler and Urs the meantime the name of the genus was Eggli (1994). „The classification of the changed to Rebutia. A workgroup of the cacti has been determined during the last International Organization for Succulent decennia to a large extent by amateurs Plant Study (IUBS-IOS) was established (persons without basic knowledge of sys- in 1984 in order to obtain consensus in the tematic botany) “. This has left deep traces classification of the Cactaceae. Some as a result of which cacti now are in the years later the CITES Plants Committee bad books of scientific botany. “However proposed that the IOS should be supervi- many interested amateurs have enlarged sor of the compilation of what was going the body of knowledge about cacti by their to be the first edition of the checklist. detailed specific knowledge of forms and Amateurs were confused. How to interpret by their fieldwork, equally often they have this list ? Some believed, it was only caused confusion by their lack of meant for the use of government customs knowledge of important biological con- and excise organisations. Or was it a list cepts (evolution, variability, biology of with obligatory names superseding all populations, and so on) and of rigorous Fig. 10: Weingartia neocumingii HS 93 Fig. 11: Sulcorebutia totorensis JK 152

Fig. 12: Sulcorebutia jolantae HS 68a Fig. 13: Sulcorebutia mentosa JD 175b working methods.“ Nyffeler and Eggli use to collect his data ? Perhaps he used argued further: “It is the task of systematic only morphological data. botany, to bring about organisation, in the David Hunt (2006) published the New huge number of plant forms according to Cactus Lexicon as a kind of honour to the postulated phylogenesis.“ Backeberg. I can imagine that the The IOS-workgroup had started with a list „splitter“ Backeberg would have felt so- of „undisputedly“ accepted genera. For mewhat displeased if he had seen the other genera there was a vote following a book. The genera Sulcorebutia and Wein- discussion. I gladly accept that all mem- gartia had been merged into Rebutia. Hunt bers had significant observations and con- argued, that Rebutia could be seen as a clusions at their disposal, which meant genus of “convenience“. It could possibly their opinions were valuable. But we do be biphyletic or polyphyletic. I have not not know these observations and conclu- the faintest idea what is meant by these sions. Is my question justified, is a sulco- two sentences. The list of names of rebutia-specialist automatically qualified species was not synonymous with the to vote about mamillaria ? By the way Cites Checklist. I did not find any expla- who supplied the (hypothetical) history of nation. In summary dogmas seem to be descent ? Which method did this person circulated and have to be followed blindly context. (2005) Can a layman have an opinion about the method of the bota- nist ? “Botany is a natural science and from this perspec- tive one can judge when scientific statements are justi- fied according to scientific criteria.“ and further on:

Fig. 14: Sulcorebutia albissima JK 39 “The basis of all natural scientific investigations are only by allegedly incompetent people like me. objective data.“ Obviously Ratz was not Even if they have real doubts about the very convinced about the basis for all sorts postulated relationships. Sometimes I of comments, concerning the (cactus) wondered if Jaap might have been a mem- taxonomy. ber of the workgroup. Some projects Also on the level of species we have come Data, immediately apparent from obser- across a similar situation. Willi Gertel ving the plant, turned out to lead to shaky (1996) for example mentioned Sulcorebu- conclusions as you can see from the above tia steinbachii ssp. tiraquensis var. to- remarks. This may be made even clearer torensis subvar. totorensis. We might hope by the next example. A plant in the green- that this was based on a profound investi- house of Karel became 15 cm high. A gation so that this combination of names cutting of this plant which I received got would conform to the phylogenesis. But I to only 10 cm. Circumstances and care could not find any information about what influence the outward appearance. exactly was investigated. By the way, Buxbaum had restricted himself to charac- could we accept a more close relationship teristics of flower and seed as these would with S. jolantae ? (Fig. 11 and fig. 12) Or always represent „primitive“ characteris- even with some populations those called S. tics, meaning that the pressure of the envi- frankiana ? Using morphological charac- ronment would have less effect on them. teristics or isoenzymes this might be pos- A further clarification : if a certain chara- sible. Would it be an argument to see for teristic is found on several plants of the example S. frankiana as a subspecies of S. whole group, it can be interpreted to be steinbachii ? I do not expect much enthu- „primitive“. It is unlikely that such a cha- siasm for amateurs putting forward such racteristic came into being many times at ideas. But they are inspired by objective different places independently of each data. other. Interesting are comments of L. Ratz in this tionship. I simply cannot un- derstand the decision to avoid the name „albissima“. Was this choice purely subjective ? All statements about rela- tionships are based on probabi- lity. Therefore all of them are hypotheses. One can wonder how to eva- luate a hypothesis when faced with evidence that in view of the hypothesis is improbable.

Meanwhile the magic word seemed to be DNA, but this was not a realistic approach for the common amateur. In 2004 I was a member of a group of enthusiastic amateurs who tried to investigate rela- Fig. 15. Cladogram according to Ritz tionships, using selected isoen- zymes. One can take the characters of One can wonder if a characteristic easily these isoenzymes to be primitive. After changes from one manifestation into ano- all, if a mutation is not favourable, there is ther and afterwards back into the original, serious chance, that the individual with creating a kind of blinking light situation. such a mutation will have died because of Let us start with a population, of which all a failing digestion, before it got descen- plants have brownish black spines. By a dants. favourable mutation this population will The result of comparing of isoenzymes obtain only white spined plants. But after showed that Sulcorebutia is less similar some generations a next mutation causes with Rebutia, but it was not really possible only brownish black spines again. Nobody to separate it from Weingartia. Would will take such a story to be probable but Brandt be declared right after all ? Some rather assume that the brownish black participants did not appreciate the results spine of the first population, for example of the project very much, perhaps because Sulcorebutia mentosa (Fig. 13), is a primi- it often did not confirm their postulated tive characteristic. Likewise the white relationships. Of course one can question spine in the other population, for example the supposition of actually knowing the Sulcorebutia albissima (Fig. 14) will be a correct outcomes in advance. If that were primitive characteristic. There is no the cases the interpretation of results may known report of a population in which well be biased to confirm such supposi- both forms appear, next to each other. In tion. Detlev Metzing rejected the results of the Cites Checklist is S. albissima taken as the project stating that the database was a synonym of S. mentosa. By the text too small. When I asked him too small for above I do not intend to deny a rela- what, he answered with an amiable smile. Fig. 16: Sulcorebutia spec. Fig. 17: Sulcorebutia cardenasiana HS 41

No, he did not know the aim, nor had he the weingartias and the sulcorebutias, but read the report (Pot 2006). not the aylosteras and mediolobivias. It Afterwards this group of amateurs happe- struck me that he showed only a part of ned to be able to initiate dna-research. the clad gram of Dr.Ritz, from which the Christiane Ritz and others (2007) publis- browningias had been removed. I don’t hed results based on chloroplast markers : know the algorithm behind this clad gram. indeed it was not possible to separate But in the cladograms prepared by me the Weingartia and Sulcorebutia. The original position of every individual is determined group of rebutias from Argentina and by the presence of all other individuals. If Browningia could be seen as a satelite a plant is removed, different arithmetical group of the weingartias. The relationship averages will be found and therefore the with Aylostera, Mediolobivia and Gymno- position of some of the plants may change. calycium was shown to be significantly For the record : the vertical distances in less strong. Therefore Rebutia minucula these cladograms have no other purpose and Aylostera pseudominuscula could not than to keep the diagram legible. Thus belong to the same genus, contrary to what they have nothing to do with the closeness had been believed for many years. (Fig. of the relationship. 15) Note, that I put alternative names, Ritz stated, that her project confirmed the used by amateurs, into the clad gram be- observation of the branched funicles by hind the ones used by Ritz. Hentzschel. The same goes for the hairless Again some people were not fully satisfied pericarps of Weingartia, Sulcorebutia and with this result. For example how can one Rebutia. Obviously the error of reconcile, that Browningia and Rebutia are Hentzschel had not been discovered and that close in the clad gram ? Did they sus- Brederoo and Donald seemed to have pect a too small database in this case as done careless observations. well ? Or were the browningia samples not In my opinion a spicy detail is the names very reliable on further consideration ? Or used for species, which would correspond would it have been better to use different to those in „Das große Kakteenlexikon“ of markers, assuming that this was possible ? Anderson in the case of the rebutias and Gordon Rowley (2009) proposed a genus the sulcorebutias and to „Die Gattung Rebutia, in which he included the original Weingartia“ of Augustin and Hentzschel rebutias, in the case of weingartias. It is reassuring Fig. 18. Hybrid of Weingartia fidana (fig. 5) with Sulcorebutia tarijensis JK242 to see that in the clad gram all canigue- not immediately recombine all sulcorebu- ralii’s, all neocumingii’s and all steinba- tias to Weingartia. I imagine it was not her chii’s are close together. The position of S. intention at all. Ritz is a geneticist. A ge- cardenasiana did surprise me. In the upper neticist is not by definition a taxonomist. part you find a S. spec. (Fig. 16). This By the way, it is not known to me if taxo- might be a very interesting plant, not only nomist is a recognised profession with a because it is not well known and wanted. protected status. Possibly everybody is Though it has been discovered at the same allowed to call him- or herself a taxono- latitude as S. tarijensis, it reminds one mist. very much of S. cardenasiana. (Fig. 17) Günter Hentzschel and Karl Augustin Suppose – which I do not suspect of (2008) published the second part of „Die course – that some error did creep in. In Gattung Weingartia“. They concluded that that case S. cardenasiana would occur in morphologically no fundamental diffe- two different clades. This I would consider rences existed between Weingartia, Sulco- very questionable. rebutia and Cintia concerning body, flo- West European amateurs of the continent wer and fruit. Also they had made several prefer to keep old, established names, attempts to hybridise, between these gene- even if they are not quite correct. It was a ra, the results of which could be consi- relief that the „cardenasiana“ mentioned in dered rather successful, while they had the clad gram is called S. langeri every hardly any successful results with other time in the known collections. genera. The consequence of these observa- Some amateurs wondered why Ritz did tions was the recombining of all sulcore- Hentzschel, because „sufficient hybrids by chance already existed in the collections“. Pip Smart showed me years ago with pride a few hybrids, caused by pollination on purpose. This had not been an easy result. Johan de Vries and I have the same expe- rience. (Fig. 18) I assume the remark about hybrids by chance can easily mislead us. The pre- sence of fruits is not necessarily evidence of hybridization. This is also well explai- ned by non hybridized pollination, which is very likely to occur as can be seen by the following. Let us accept that on a table in the garden there are 100 plants, which all have one and only one partner for ferti- lization. Thus on this table exist 50 pos- sible matches. Suppose all plants bloom at the same time with only one flower, and every flower is equally attractive to a bumblebee. The insect has landed on a flower. After some time it will ascend and land again. For every flower the chance Fig. 19. Cladogram compiled by CactusData that it will land on it is 1 of 100, or 1%. So there is a chance of 99% = 0,99 that the butias to Weingartia, which had partially flower of the match plant is not visited. been done already by Brandt, about 30 After a second landing this chance will be years before. However some oddities 0,99 × 0,99 = 0,992. After a number n struck me. At page 771 hybrids are shown landings the chance has become 0,99n. If of S. canigueralii and W. neocumingii (fa. n = 459, the chance of no visit has become sucrensis). 14 km from Sucre, along the less than 1%. So while it may be true that road to Aiquile, sulcorebutias and wein- all sulcorebutias are strongly related, this gartias, usually called canigueralii and suggestion has not been confirmed by neocumingii (fa. sucrensis) as well, grow observations of fertilization by chance. next to each other. A natural hybrid was There is still another consideration. If never reported from that place. Should I hybridization in nature were that easy, one believe now, that these plants somehow would not expect so many strongly diffe- got the wrong names? Or have canigue- ring populations so close together. The ralii-populations different characteristics, question of relationships motivated me to by which some will yet hybridise easily create a database, which at the moment with W. neocumingii and others will not at contains almost 50 characteristics of 1700 all ? Hybridising of different species of individual plants of the genus Weingartia. Weingartia with each other or Sulcorebu- Which characteristics are indeed signifi- tia with each other was not tested by cant and which are not ? From the above we’ve learned that experts often disagree. result as encouraging. Perhaps the following leads to a sensible The key to recognition (Pot 2009) was choice. A characteristic which is fairly based on this database. Last year somebo- constant for the whole population can be dy discovered, that in this key Sulcorebu- used for this population, but perhaps not tia steinbachii was recognized in different for another one. Sulcorebutia mentosa for ways, in which moreover the appearance example has magenta flowers, never red or of the populations were different. Actually yellow. Thus the colour of the flower is a he had chosen to look for all steinbachii’s significant characteristic in this case. In in one search operation, like in the New different populations, which usually called Cactus Lexicon for example. S. losenickyana all sorts of colours occur. According to the NGL a steinbachii is For such populations this characteristic recognized by: „Body very variable, clus- will not be of much value. tering; stems depressed-globose; tubercles Now I selected plants, similar to the plants in circa 13 spirals; central spines 1-3,

position of radial spines distance lower filaments to shape of scales on the tube pericarp / length pistil colour of radial spines length of filament colour perianth in top structure of surface of radials distance anther to edge colour perianth above upper perianth insertions

length radials distance stigma to perianth colour stigma position central spines diameter of stigma colour filament shape of the plant body angle tube below insertions colour throat

length of the pistil shape of perianth colour scales on the tube area of the insertions of edge of perianth colour scales on the pericarp filaments How to continue ? How will somebody be able to draw con- Last summer I was visited by a very se- clusions about the range of variability of rious amateur. He confided to me with a these undefined units ? Who will be able serious look, that he would accept a classi- to draw conclusions about strong rela- fication in future only if it had been fully tionships between such units without defi- based on dna-research. Of course I appre- ned ranges of variability ? Who will be ciate this very much. But in the meantime able to understand that nevertheless, nu- I’m a little worried. Until now I have not merous relationships are conjured out of a heard of any research that will give solu- hat ? tions at the level of species. But suppose it I am not a taxonomist. But it is my im- does happen. Will we, amateurs, be able to pression, that empirical experts of the do something with it ? Will relationships previous century could not have suspected be confirmed by easily observable charac- for a moment the variety of forms of some teristics ? plant families. People have found during Basically they are trying to construct a the last decennia unexpectedly many po- system based on relationships. But at the pulations of Weingartia, of which the ma- same time we want to know what we are jority is well distinguishable from all other speaking about. DNA-research will be a populations. Nobody – with or without a „black box“ to most of the cactus ama- definition of the concept „species“– will teurs. A solid system, which however is expect to recognize a thousand species. not understood because of obscure back- But also nobody is able at the moment to grounds has little sense in amateur circles. create an understandable and acceptable The same goes for a classification based survey of relationships in this genus. Of on carefully hidden morphological obser- course one can ignore this multitude of vations. forms. But it does not make any sense to Maybe Karel was close to the truth when dispose of all unknown populations as he said: „The only thing that really matters „flowerpot species“. The same goes for is that you understand me. Isn’t that accepting all relationships blindly without right ?” Suppose somebody speaks about supporting data. for example Sulcorebutia tarabucoensis, Moreover now we are dealing with new which has according to the first descrip- modern technology, which may put postu- tion few brown radial spines, about 8 ribs lated relationships in a very different light. and red-yellow flowers. But he means a I can easily imagine a Gordian cactus plant with many white radial spines, about knot, which we will not be able to un- 13 ribs and a magenta flower, from a very tangle by continuing as we have in the different population. My reaction would past. Yet I keep hoping for new ap- be, this person does not know Sulcorebu- proaches, which will lead us out of the tia tarabucoensis. He for his part perhaps dilemma. Maybe one day the knot will be believes that I do not understand the range untangled in a more sophisticated way of variability of this species. Thus we will than used by the legendary Alexander. But not understand each other. some splitting may be not avoidable. Professionals have not succeeded in cons- People like Nyffeler and Eggli may find tructing an unequivocal, generally ac- my considerations anarchistic. This howe- cepted definition of the concept „species“. ver is not my aim. I would welcome an explanation by an expert about the whole Gertel, W. (1996): Sulcorebutia tiraquen- area in clear and workable terms. Maybe sis, Kakt. and. Sukk. 47 (6): 132-139. this paper will give the initial impetus. Hentzschel, G. (1999): Het geslacht Sul- corebutia Backeberg emend. Hentzschel, I like to thank Jim Gras for proofreading Succulenta 78(3): 131-142. the English text. Hunt, D. (1999): Cites Cactaceae Check- list, pp. 126-137, 143-145, Kew.

Hunt, D. (2006): The New Cactus Lexi- Literature : con, p. 248, dh books. Pot, J. (2006): The use of isoenzyme anal- Augustin, K. & G. Hentzschel (2008): ysis to study the dispersal of Sulcorebutia, Die Gattung Weingartia, Gymnocalycium Cactus & Co 4(10): 235-250. 21(2):767-782. Pot, J. (2009): Sleutel voor planten van Backeberg, C. (1959): Sulcorebutia, Die het geslacht Weingartia (Sulcorebutia), Cactaceae, Band 3, p. 1555, VEB Gustav Succulenta 88(3): 132-138. Fischer Verlag, Jena. Ratz, L. (2005): Betrachtung zum Pro- Backeberg, C. (1933): Echinocactus blem Rebutia pygmaea-Rebutia haagei, Fidaianus, Der Kakteenfreund: 90. Echinopseen 2(l): 1-5. Backeberg, C. (1951): Sulcorebutia, no- Ritter, F. (1980): Kakteen in Südamerika, vum genus Backbg. Cact. Succ. J. (G.B.) Band 2, pp. 641-642, eigen uitgave, Span- 13(4): 96, 103. genberg. Berger, A. (1929): Kakteen, p. 196-197, Ritz, C. et al. (2007): The molecular phy- Ulmer Verlag, Stuttgart. logeny of Rebutia (Cactaceae) and its Boom, K. (1962): De correcte naam van allies demonstrates the influence of pale- Echinocactus cummingii Salm-Dyck, ography on the evolution of South Ameri- Succulenta 41(9): 115-118. can mountain cacti, Am. J. Bot. 94(8): Brandt, F. (1977): Weingartia oder Sul- 1321-1332. corebutia!, Kakteen- und Orchideenrund- Rowley, G. (2009): Rebutia reappraised, schau 5: 68-70. Cactus World 27(2): 88-90. Brederoo, A.J. & J. D. Donald (1972): Salm-Dyck, J. (1849): Echinocactus cum- Sulcorebutia Bckbg emend. Brederoo et mingii, Cact. Hort. Dyck., p. 174. Donald, Succulenta 51 (9): 169-174. Schumann, K. (1895): Rebutia minuscula Brederoo A.J. & J. D. Donald (1981): K. Sch. Eine neue Gattung der Kakteen, Blütenuntersuchungen bei Weingartia und MfK. 5: 102-105. Sulcorebutia, Kakt. and. Sukk. 32(11): 270 Spegazzini, C. (1923): Breves Not. Cac- -273. tol., An. Soc. Cient. Argent., XCVI: 75. Buxbaum, F. (1967): Der gegenwärtige Werdermann, E. (1931): Neue Kakteen Stand der stammesgeschichtlichen Erfor- im Botanischen Garten Berlin-Dahlem, schung der Kakteen, Kakt. and. Sukk. 18 Notizblatt des Botanischen Gartens und (1): 3-9, 18(2): 22-27. Museums zu Berlin-Dahlem, Nr. 104. Bd Donald, J. D. (1971): In Defense of XI. Sulcorebutia, Cact. Succ. J. (US) 43(1): 36 Werdermann, E. (1937): Beiträge zur -40. Nomenklatur. 4. Weingartia Werd., Donald, J. D. (1980): Probleme bei der (Spegazzinia Bckbg. non Saccardo 1886!), Trennung von Sulcorebutia und Weingar- Kakteenkunde (2): 20-21. tia, Kakt. and. Sukk. 31 (11): 321-327 Eggli, U. & R. Nyffeler (1994): Die Kon- sensusklassifikation der Kakteen, Kakt. and. Sukk. 45(7) :142-149. This article originally written in Dutch by Johan Pot was published in the journal Succulenta 90:5 (2011) (p. 227 -237) Until today, this English version (translation by Johan Pot) was unpublished.

Published with the permission of the author.