Printed in British Cactus and Succulent Journal 4:45 (1987).

Haworthia is indeed a popular genus which seems to inspire a great deal of controversy and confusion. One would have to be very thick-skinned to be able to ignore past history and not plead for forgiveness for similar transgression. I was just busy trying to clarify, in my own clouded mind, the problem of H. pumila (L.) Duval, when I saw Will Tjaden’s little article on the subject in this journal (3:88, 1985). Gordon Rowley in the same issue reviews the recent books on Haworthia and also mentions the H. pumila versus H. margaritifera debate. Coming so soon on the heels of Fearn versus Cole and Walker versus Bruyns, it would be insensible for Bayer to take up the cudgels against anyone.

In any case I frankly do not know what to do about the problem of the name-game so well expressed by both Rowley and Tjaden, and yet I shamefully have to admit my displeasure at their contribution, or lack of it. In the case of Tjaden, I agree with his comments on name changes and respect this view far more than he suggests. My distress at the recognition of H. pumila (L). Duval is greater than Tjaden has conceived, and all the more because I knew that Col. Scott’s solution offered in 1978 was not correct. Col. Scott was assisted in this instance by Dr L.E. Codd, who is one of our most respected taxonomists. Unfortunately they overlooked Burmann’s Flora Capensis of 1869 and also the fact that another species (H. minima) was involved.  While I accepted their decision in the interests of stability and peace, Dr. Onno Wijnands pursued the matter a little more vigorously.

The intention of the Code of Botanical Nomenclature, to bring order and stability to the names of plants, seems fairly obvious. It also seems to me that the Code, being as complex as it is, is either an ass or can often be made an ass of, because its various articles can be used in contradiction to one another. For Mr. Tjaden’s benefit particularly, let me just quote, article 55.2 which is directly relevant to H. pumila… ”On transference of a specific epithet under another generic name, the resulting combination must be retained for the species to which the type of the basionym belongs and attributed to the author who first published it, even though it may have been erroneously applied to a different species.” One of my first experiences with the Code was in the presence of three distinguished taxonomists (one professor and two PhDs) trying to get guidance on the question of application of Linnaeus’s epithet pumila. The problem was so skillfully compounded and evaded that I felt that this simple mortal would have to stick to easier things. I really thought that Dr. Codd had helped solve a complex problem and I am as disconcerted as anyone else that Dr. Wijnands has established otherwise, arguing:—

1. H. pumila (Alt.) Duval is a homotypic synonym of H. herbacea (Mill.) Stear.

2. H. pumila (L.) Duval does not exist, and H. pumila (L.) … cannot be taken up because it would be a later homonym of H. pumila (Ait.) Duval.

3. H. margaritifera (L) Haw. is the correct name for the species concerned.

Gordon Rowley is right in surmising that we have not yet heard the end of the story. I do not yet know the correct position and can only say confidently that it has not yet been arrived at. (A whole lot more straw in the wind followed. The actual outcome was the typification of the Linnaean epithet pumila on as the Commel illustration t10 and this is the valid name in Haworthia). Tjaden is quite wrong in referring to a name-game being played by botanists. Names are part and parcel of the communication process and confusion can just as often be put at the door of gardeners as anywhere else. Duval was a gardener.

Regarding Gordon Rowley’s review, there are a host of major and minor discrepancies between Col. Scott’s perceptions and my own, and I am particularly disappointed in that the review examines none of them nor even reveals them. One of my chief complaints about reaction and response to even my own writing is the undiscriminating attitudes of the audience. Even my respect for Gordon Rowley does not deaden the impression that he has not actually read and comprehended what either Pilbeam, Scott or myself has had to say. Without meaning to be unkind or offensive, I think a much better review could have been written unless Gordon Rowley also finds the audience unselective.

Regarding Rowley’s references to monographs and revisions, I would like to point out that I have used neither term for my work. The cytological data which Rowley refers to as a vast body of published research is an indictment of the subject rather than a source of worthwhile information. Riley and Majumdar (1979), and here I think it unlikely that Rowley read the book, state “all determinations of chromosome numbers should be regarded with suspicion” and this sentence is qualified by a reference to “if they were made from botanical gardens collections of long standing”. It would have been better to have said “from plants of doubtful authenticity”. A common problem with cytological work is that one has to be sure of the material worked on, and although H.P. Riley does cite voucher specimens (which are lodged here at the Compton herbarium), the origin and identification of the material remains doubtful. Even a good dry specimen is not easy to identify. At present the best cytological work, in the sense that Rowley asks for, is that being published by Dr. Canio Vosa at Oxford. At least emphasis is placed on authenticity of material. Dr. Peter Brandham at Kew has also published proper cytological work on Haworthia and all the references available to me are listed in my handbooks. I am sorry if it is not obvious that I had also considered the contents. Brandham’s work is more orientated to chromosomal aberration than to species resolution, and in any case the results of his work on H. coarctata and H. reinwardltii seem to substantiate my mass transfer of varieties.

Rowley for some odd reason cites the basic chromosome counts in H. venosa, H. recurva and H. tessellata. These are all 14 but with polyploids in the latter two. What this bit of information is supposed to convey to anyone evades me completely. Scott mentions these counts too in Cactus and Succ. Journ. (U.S.) 50:77 (1978) with the same degree of breathless reverence. At least Scott points out that there is little possibility of knowing just what Ferguson in 1926 was naming H. recurva. If there is anything in the literature (including the work on chromatography by Riley and IsbeIl, 1963) to suggest any arrangement of species substantially different from mine, I stand helplessly to correction.  My handbook was not written to duplicate what I wrote in National Cactus and Succulent Journal 28:80 (1973).

Regarding leaf surfaces, Dr. David Cutler was kind enough to photograph a good number of leaves using electron scanning micrography. I wonder if anyone can read more taxonomic sense into these pictures than I have been able to extract from them. It seems only possible to interpret them against a background of origin and with some prejudgement of taxonomic order. Dr. M. Hayashi of the Tokyo Breeding Research Institute has some very valuable results from tissue culturing and this will to some extent improve our understanding of species relationships. But any method like this which departs from standard herbarium practice of comparing dry barely recognisable specimens, moves onto shaky ground because of the real limits to the material that can be properly examined. At some time or another one has to reach some kind of a decision, and if we are going to wait for clear-cut anatomical, cytological and physiological studies, will we ever get revisions, let alone monographs?

The appearance of Col. Scott’s revision of Haworthia does in fact pose many problems for me and a superficial review is not very helpful to anyone. Who in his right mind wants to enter into a needless and fruitless disputation about who said what when, and about who is right and who is wrong? On the other hand, concerned readers must surely want to know what skeletons are lurking in the cupboard – so let’s open the door and see.

H. herbacea (Mill.) Stear is typified by an illustration in Boerhaave’s Index Alter Plantarum (1711). I apply this name to an element which Scott calls H. arachnoidea, in turn typified by an illustration in Commelin’s Praeludia Botanica (1703). Wijnands points out that this illustration is actually in better agreement with Scott’s H. setata than his H. arachnoidea, and further confounds the issue by stating that H. arachnoidea sensu Bayer is the same element as H. herbacea sensu Bayer, but omitting this relevant citation from his synonymy. Setting this aside, I would concede that my idea of the Boerhaave illustration representing my species causes problems, as I have long nursed the secret apprehension that the illustration more nearly represents H. magnifica V. Poelln. In a way this is a problem that both Scott and myself have been advised to avoid, i.e. selecting neotypes which may later be upended, it would have been difficult for me in 1976 to predict what interpretation Scott would adopt in 1985 and Tjaden may place the blame for confusion where he wishes.

In the case of H. angustifolia Haw. there does not seem to be a problem because nearly all authors imply that Salm Dyck’s illustration (Monogr. 13, t.2 1836) of Aloe stenophylla Schultes represents the species. Fourcade (in Trans. Royal Soc. S. Afr. 21:78, 1932) gives a new name for H. angustifolia sensu Baker, viz. H. monticola. He cites A. stenophylla Schultes and also Salm-Dyck t.2 as synonyms, and states expressly that Schultes’ epithet cannot be taken up in Haworthia because of the priority of H. stenophylla Baker (now Chortolirion). Schultes’ A. stenophylla is a direct transcription of H. angustifolia as described by Haworth. Under Article 63 of the Code, H. monticola is a superfluous name for H. angustifolia Haw. because Fourcade has indisputably included the type of that species.  In analysing the discussion of his H. monticola (as I pointed out in Nat. Cact. Succ. J. 37:31, 1982 – one of many references not cited by Scott) it is obvious that Fourcade wholly misinterpreted both the Haworth and Baker descriptions. I do not contest that Fourcade had a different species, but this is only evident from mention of the origin of his specimens. He describes a plant which has translucence in the lower leaf. This is quite nonsensical and there is no comparision between the chlorosis of buried leaves and the translucence of exposed leaves in those species which do exhibit translucence. Scott’s description under H. monticola in this respect also deviates from Haworth, Baker and Fourcade, because he writes of a leaf with a semipellucid upper surface with a number of pellucid spots in the upper third. This is not at all evident in Scott’s illustration of the species and I would go so far as to say that he has two different elements in the picture. The foreground and middle-right plants are in my opinion H. chloracantha var. subglauca (V. Poelln.) Bayer. The other plants resemble specimens I have of H. divergens Bayer from Uniondale which are relatively unspotted. I did not take up Fourcade’s name for the very reasons then suspected and, even if only for that reason, see none to change now.  Therefore:

H. divergens Bayer, Haworthia Handbook :38 (1976). Type: Bayer 175 (NBG, holo).

[Syn. H. monticola Fourcade, nom. illegit., sensu Scott, The Genus Haworthia, a Taxonomic Revision, :57 (1976).]

(For the sake of peace and with some doubt about my own nomenclatural expertise, I later just abandoned any argument and reverted to the use of monticola.)

Col. Scott takes up a number of old Kew illustrations as types of Haworth’s species e.g. H. altilinea, H. mucronata and H. aristata. I put my case clearly in both editions of my Handbook and, contrary to Rowley’s supposition, these illustrations have been available to researchers outside Europe and were also seen by both Uitewaal and Von Poellnitz. Scott simply confirms the total confusion in the historical application of these names by illustrating H. habdomadis var. inconfluens (v. Poelln.) Bayer and H. cooperi Baker from Ladismith and Grahamstown respectively, as H. mucronata Haworth. His distribution map implies the inclusion of H. bolusii var. blackbeardiana (v. Poelln.) as well.

Col. Scott’s abandonment of subgenera and the characters which support them is inconsistent with exclusion of Astroloba, Poellnitzia and Chortolirion from Haworthia. If inflorescence characters of this degree are ignored identification must accordingly be complicated. I simply cannot understand Rowley highlighting practically useless cytological evidence and passing this over. The consequences to Scott’s perceptions of Haworthia are immediately disastrous where on page xii of his ‘revision’ the illustration C478 is identified as H. marginata which it cannot possibly be. Jaquin has an identical illustration in Hort. Schoenbr. :421 entitled Aloe pumilio which Scott at least has correctly labelled H. reticulata. On page xiii, C481 is obviously H. marginata and the flower positively excludes H. venosa as Scott suggests.  C313 on page xii also compares very unfavourably with H. heidelbergensis on page 131 and I find it very difficult to believe that the species is correctly illustrated anyway. These misidentifications of C478 and C481 cross the boundaries of three major groups of Haworthia first indicated by Uitewaal. Scott finds it difficult to understand why H. bruynsii and H. springbokvlakensis should be separated taxonomically when they are separated… “geographically by only a few kilometres of geologically uniform undulating ground”. I find it equally difficult to accept a classification that for example separates H. scabra Haw. and H. tuberculata v. Poelln even at species level, let alone at sectional level. Scott’s exposition of his species concept does not leave me any the wiser, and the short section on the generic concept speaks for itself.

Without wanting to be unkind, I would have thought that some of the misstatements in Scott’s book should have been apparent to a competent reviewer. Gordon Rowley, having seen H. graminifolia in the field himself, should of all people know that it does sucker. I have avoided giving habitat data like the plague because it is so difficult to make true and usable statements. One has only to read Scott’s discussion of H. pumila to see on what quicksands one treads in this respect. Read first the discussion and then refer to the distribution map. Being right here on my doorstep, I should know this species very well.  H. pumila is not found in strictly fynbos vegetation and the statement… “found under fynbos or other karoo scrub” can be valuable information only to the barmy. This species grows here on Malmesbury shales and it is only common in the Pteronia paniculata communities (one of seven such) where pH can be below pH 4 (KCI). These soils are also characterised by Euphorbia mauritanica communities where pH is often above 8, and H. pumila can be at home there too. It also occurs on soils derived from Witteberg quartzites, Dwyka tillite and Bokkeveld shales and I do not believe that this information is going to help any grower in the slightest. H. reticulata is similarly described by Scott as a component of fynbos, but is also said to grow on outcrops of shale. If anybody can be said to deserve credit for making such mutually exclusive statements then truly we can identify this epoch as the age of untruth.

Of what value is a statement for H. minima Haw. which reads… “A constituent of fynbos . . under renosterbos”?  Fynbos is one veld-type and renosterbos is a major element of another. H. herbacea (H. arachnoidea sensu Scott) is also on my doorstep, and I am more than surprised to learn that it grows in sandy, well-drained acid soils derived from sandstone. While it has the same distribution range as H. pumila it is fortunately blessed in that it enjoys a higher rainfall maximum. Who will point out that H. angustifolia var. baylissii is represented by a single clone from one of the densest populations of H. angustifolia that I have seen and that in fact has no distribution? Who will point out that Scott’s illustration of H. serrata is definitely not that of my species and that it can best be equated with the form of H. arachnoidea at the northern end of the Tradouw Pass? In the case of H. marumiana sensu Scott, he cites H. archeri Bayer as a synonym and also H. marumiana sensu Bayer. He then goes to great pains in the discussion to exclude the latter and suggest that its “systematic position is not yet clear.” Uitewaal does describe his species as freely proliferous but Scott writes… “H marumiana sensu Uitewaal” is not.

To evaluate this revision of Scott’s we should perhaps look at his section Retusae which he first revised in Aloe 11, 4:17 (1973). The synoptic key (Rowley) consists of 8 steps each with two to three species. One of these steps reads “leaves in 5-tiers” and includes H. retusa and H. heidelbergensis, while another reads leaves in 2-4 tiers” and includes H. mirabilis and H. asperula sensu Scott. Under the description of H. retusa these tiers are described as vertical, while it is wholly unclear what 2-4 tiers means in H. mirabilis or H. asperu!a. The illustration of H. mirabilis shows a plant with 8 near-vertical tiers.  In the illustration of H. asperu!a it is possible to resolve the two primary counter spirals, the three secondary spirals, and the five secondary counter spirals. The illustration of H. magnifica facing H. asperula shows beautifully the five secondary counter spirals in near vertical position. (I wrote the article in respect of leaf spirals in Aloe precisely to forestall and nudge both Dr Codd and Col Scott to a rational view.)

In 1973 Scott upheld the name H. nitidula v. Poelln. but the 1985 version does not cite this interpretation under H. longibracteata Smith where it belongs by virtue of its accompanying illustration. H. willowmorensis v. Poelln. sensu Scott 1973 becomes H. correcta v. Poelln. but is cited under the synonomy of H. mirabilis. Similarly H. sublimpidula v. Poelln. of 1973 becomes H. heidelbergensis Smith of 1985, but is not cited as such. The illustration and guide to H. heide!bergensis do not agree with the original description nor with this species as I know it.

The illustration of H. dekenahii is not that element but that of H. turgida var. pallidifolia Smith, which Scott does away with in a new concept of H. retusa.  While H. asperula sensu Scott covers everything from H. pygmaea in the east to H. pubescens in the west, four species viz. H. geraldii, H. multilineata, H. fouchei and H. longibracteata are required for a relatively simple situation around Riversdale.

These are only some of many problems which I see in Col. Scott’s book. John Pilbeam’s book is in another class. Except where he enters the taxonomic fray it is the most useful book for the collector. The most serious objection that I have to it is the chapter on geographic distribution. The only nice thing I have to say about that is Pilbeam’s kindness in concealing the origin of most of it. This chapter should have been updated and corrected before publication.

The ‘definitive monograph’ is Gordon Rowley’s pie- in-the-sky. The realities of plant taxonomy are going to prove far more disparaging of hybrids and cultivars by taking them properly out of the rank of genus and species. Haworthia has just been fortunate in having being treated by non-botanists who are sympathetic to the collector. A truly botanical dispensation could reduce the Hexangulares to 11 species, the subgenus Haworthia to 18 and Robustipedunculares to 4. Teasingly I would like to say to Gordon Rowley that he should look at more than just the dust cover when he next reviews a book.

GORDON ROWLEY comments: ”In some books the dust covers are the best parts. But to be able to state that the only reference to chromosomes in all three books is on p.73 of Bayer’s surely indicates some attempt to look between the covers, doesn’t it? Maybe I just wasn’t reading the same volumes that Bruce has. Incidentally, I would love to know what is to become of the luckless hybrids and cultivars if they are to be deprived of the rank of genus and species!

Now, perhaps, could we invite Col. Scott to round off the review with his side of the story?” ♦