Haworthia Revisited – 61. Haworthia pumila

61. Haworthia pumila (L.) M.B.Bayer comb. nov. Aloe 16: 44 (1978). Bayer: 82 (1982). Scott: 13 (1985). Type: Comm. Hort. Amstel.: 19, t 10, (1701). Epitype (ex B&M): Karoo Garden, Worcester, Compton 78963 (NBG): Aloe pumila var. margaritifera L., pro parte, Spec. Plant.: 322 (1753). L., Spec. Plant. 2: 460 (1762). Aloe pumila (L.) Burm.(f.), Prod. Fl. Cap.: 10 (1768). Thunb., Diss.: 183 (1785). Haworthia maxima (Haw.) Duval Hort. Allene.: 7 (1809). Haw., Syn. Pl. Succ.: 92 (1812). Haw., Suppl. Pl. Succ.: 53 (1819). Aloe margaritifera var. maxima Haw., Trans. Linn. Soc.: 7: 11 (1804). A. semimargaritifera var. maxima (Haw.) Salm-Dyck, Verz. Art. Aloe: 6 (1817). H. margaritifera var. semimargaritifera (Salm-Dyck) Baker, J. Linn. Soc. Bot. 204 (1880). Type: Hort. Dyck. not preserved: Aloe margaritifera (L). Miller, Gard. Diet.: 18 (1768). Lam., Encycl.: 88 (1783). Ait., Hort. Kew. 1: 468 (1789). H. margaritifera sensu M.B.Bayer, Cact. Succ. J. (US) 43: 157 (1971). Bayer: 133 (1976).: H. semiglabrata Haw., Suppl. Pl. Succ.: 55 (1819). Neotype (B&M): Salm-Dyck, Aloes Mesembr.: 6: t8 (1837): H. papillosa (Salm-Dyck) Haw., Suppl. Pl. Succ.: 58 (1819). Haw. Revis.: 55 (1821). A. papillosa Salm-Dyck, Verz. Art. Aloe.: 6 (1817). Salm-Dyck, Aloes Mesembr. 6: 14 (1836). Type: Salm-Dyck, Monogr. 6:t4.

pumila: dwarf.

(Note: non H. pumila (L.) Duval, Hort. Alene. 7 (1809). Non Aloe pumila (L.) Haw., Trans. Linn. Soc.: 7: 10 (1804). A. arachnoidea var. pumila Ait., Hort. Kew.: 468 (1789). Aloe arachnoidea var. pumila Willd., Spec. Plant.: 188 (1799). Typified by: Boerh., Index Pl. Lugdb. 2: t131,(1720). = H. herbacea (Mill.) Stearn.).

Rosette stemless, slowly proliferous, to 250 mm tall. Leaves to 140 X 20 mm, nearly as thick as wide, attenuate, spreading, lanceolate-deltoid, surfaces scabrid with raised, rounded non-confluent tubercles, colour brownish- to olive-green. Inflorescence sparsely branched, lax. Flowers tepals fused, tube straight, lobes abbreviated, veins brownish-green.

1982 – Linnaeus is regarded as the founder of modern botanical nomenclature, and the example of H. pumila shows what confusion reigns where he did not give a clear lead.  C.L. Scott correctly resurrects the name H. pumila, although Duval had a quite different species in mind.  The confusion that resulted from Salm‑Dyck’s publication Catalogue Raisonne des especes d’ aloes (1817) has simply worsened to the present day.  This is so much the case that some of the synonyms are regarded as separate species.  H. pumila is the largest of the haworthias growing up to 300mm tall.  It is usually dark brownish‑green with large off‑white rough tubercles.  The tips of the florets are also brownish‑green.  Its distribution is primarily the Karroid Broken Veld of the Worcester/Robertson Karoo, but it occurs in the Hex River Valley, Montagu, Barrydale (unconfirmed collections) and Stormsvlei areas.  The plethora of names is not justified by the variability of the species.  Only once has a completely glabrous form ever been collected (undocumented), and it is only at Lemoenpoort south of Worcester where the plants are sparsely tubercled.  H. papillosa has only been upheld on the basis of leafy stem and regular arrangement of the leaves in vertical tiers.  This is nonsensical, as the leaves are arranged in spirals and inevitably some specimens will appear to have vertical leaf tiers (see Bayer, 1973).  A problem arises where H. marginata and H. pumila grow together as happens at Bonnievale, Ashton and Drew, because, improbable as it may seem, the two species appear to intergrade.  H. pumila is strictly a winter‑growing species.  It is seldom proliferous but clumps do occur and it can be propagated from whole leaves. In cultivation it is also sensitive to leaf aphis in the heart of the plant.

1999 – Many experienced and skilled taxonomists have given attention to the nomenclature of this species, and failed to have produced a good answer.  Part of the problem is that the types upon which the names have been based have simply not been understood.  H. pumila (Haw.) Duv. is based on the same type as H. herbacea (Mill.) Stearn, and it appears to me that it is therefore unavailable for this species.  There are actually six original illustrations to serve as types, and these are set out in the synonymies of this species and of H. minima.  If these are followed through properly, it is apparent that the name H. margaritifera Haw. follows Aiton and is based on an illustration that is also H. minima.  Haworth’s H. margaritifera is synonymous with his Aloe margaritifera var. major of 1804 which specifically excludes t10 of Commelin.  It is based on the illustration by Bradley (t21, 1725) and no one can dispute that Bradley’s description is of the smaller species (“leaves…. a little more than one inch in length”).  H. margaritifera is therefore a later synonym of H. minima.  Five of the six illustrations available as types are, in my opinion, H. minima.  This also reflects two things.  Firstly, H. minima would, by virtue of its distribution south and east of the Langeberg mountains, have been far more exposed to early exploration and travel.  Secondly it is far more amenable to cultivation than the more difficult to grow H. pumila.  The fact that competent taxonomists have failed to reach consensus and an understanding of this species leads me to unashamedly to present my own solution.  What Haworth may have written about consummate skill and unwearied attention, also seems to apply to the rules on nomenclature.  To survive in that environment seems to call for the legal mind of a judge and a minimal knowledge of the plants one is attempting to designate.

The distribution of H. pumila really is interesting as it is found in several places in the western Little Karoo, from as far north-east as the Anysberg Reserve.  It is also common in the area around Pieter Meintjies and Matjesfontein, west of Laingsburg, impinging on the Great Karoo.  The record from west of Barrydale requires confirmation.  It hybridises with Astroloba muricata near Montagu.  E. Aslander has found the same hybrid to the east of Barrydale but only observed the Astroloba.  Possibly the other parent may be H. minima.

Distribution: 3319 (Worcester): Osplaas, DeDoorns (-BC), Bayer (NBG); Worcester (-CB), Bolus (BOL); Karoo Garden (‑CB), Compton 18963 (NBG), Smith 7202 (NBG), Barker 5097 (NBG); Doornrivier (-CD), deVilliers in NBG477/36; Lemoenpoort (-CD), Bayer (NBG); Mowers (-DA), Bayer 6622 (NBG); W. Rooiberg (-DC), Bayer 6623 (NBG); Vrolijkheid (-DD), Burgers 58 (PRE); Robertson (-DD), Marloth 7982 (PRE).  3320 (Montagu): W. Matjesfontein (-AB), C.A. Smith 2508a (PRE); Jagersfontein (-AB), McKenzie (NBG); Matjesfontein (-BA), Pillans 875 (BOL); Anysberg (-BC) (NBG); Baden (-CA), Neil (BOL); near Montagu (-CB), Esterhuysen (BOL), Bolus 7513 (BOL); Robertson (-CC), Hurling & Neil 13735 (PRE); W. Barrydale (-DC), Smith 7127 (NBG).  3420(Bredasdorp): Stormsvlei (-AA), (NBG).

Inadequately located: Ex hort, Peers (NBG); Ex hort Whitehill (NBG); Cape, Marloth 5962 (PRE);

Aloe pumila, Haworthia pumila; what or who is confused?

With some astonishment I learned recently of an opinion that the name pumila as used in Haworthia is confused and therefore should be discarded. There is no doubt that it is difficult to unravel the literature and the usages of the name, but I think there are a number of important questions which should first be answered. Is the name confused or is it us and others, as individuals, who are confused? It is worth asking such questions because names are words in the process of communication and mutual understanding around which all knowledge and its sharing revolves.

If the facts of the matter are properly examined there is a clear path of events. It may be very complex and take many words to explain, but it is there. If a reason has to be sought for confusion I have no doubt that it can be tracked to the door of the International Code of Botanical Nomenclature. My reasons for saying so should become clear.

Paul Forster, who is himself a professional and highly qualified taxonomist, commented on my treatment of the name H. pumila as follows “We do not live in a vacuum”. Here he suggested that I had failed to consult authority and that there was professional opinion available that would have resolved my confusion.

If one thinks about Paul Foster’s statement it is quite evident that:
1. he does not consider himself able to resolve the problem (this is a valid point because to do so does require knowledge of the elements involved, and also requires insights into the intricate mechanisms of the Code which are beyond most).
2. he simply believes that there is someone in the hierarchy who is competent and able to do so.
3. I had not thought there was.

The latter point is simply not true as I have consulted almost every single botanist and non-botanist (yes, many non-botanists are free to dabble in plant taxonomy because of the non-scientific nature of the activity) available to me. A list of names and also authors and published references would fill a few pages.

Turning to the name “pumila” itself, it can be shown very clearly (botanical terminology to mean something quite the opposite, but which I use here in the true sense of understandable) that Linnaeus used the name for four “varieties” of small Aloe. Confusion certainly followed about which of his names should be used, but there should be no doubt about which four these were.

The problem arises out of the Code and how, relating to names for those four “species”, it is now interpreted and by whom. It becomes a juristic problem requiring a great deal of fancy intellectual footwork, complicated by an approach to the code which largely ignores the basic intention to bring stability and uniformity to plant names. It also involves use of words like “validity”, “legitimacy”, “intention”, “relevance”, “strict interpretation” and “according to article…” with gay abandon.

Without going into the very extensive historical detail which includes the fact that the names margaritifera and pumila were used for the same single thing, it is almost sufficient to say that Col. C.L. Scott took the plunge. Defying the predictable response, he took the assistance of Dr L.E. Codd, who was then and still remains a highly respected taxonomists. Together they concluded that because of what and despite what, anyone had done with those two names, the name “pumila” was valid and correct for the species represented by the illustration t10 of Commelin 1701.

They argued that despite the fact that the name had already been used in Haworthia for another species, this did not preclude its correct use for the one represented by the Commelin illustration. The fact that confusion may have arisen and continues to this day, is to my mind correctly laid at the door of Dr W.T. Stearn, paragon and patron saint of plant taxonomy.  He wrote a paper in 1938 in which he annotated the names of Salm-Dyck and reconciled them with Berger’s revision of Aloe in 1908.  What he did was also to  typify the name Haworthia herbacea (Miller) Stearn. In my opinion, had he known anything about the actual species involved, he would have recognised the problem arising from the inclusion of “pumila” in the history of that name and resolved it accordingly.  He did not.

After Scott (and Dr Codd) had reached a decision in 1978, I still remained in some doubt using the name “pumila” until forced into a decision when I undertook to write a revision of the genus myself. I had been asked to write a synopsis of the species of Haworthia occurring in the Cape Floral region.  This request was by the taxonomists Dr. J. Manning and Dr. P. Goldblatt and I in turn asked them (as had become my routine practise with taxonomists) to clarify the use of the name “pumila”. They contacted Dr Fred Barrie who I presume is a figure in the hierarchy alluded to by Paul Foster, and the following was his reply. “According to the Linnaean Typification Project database, Aloe pumila var. pumila was lectotypified by Wijnands (Taxon 34:310, 1985) on Commelin, Hort.Med.Amstel.2:t.10, 1701. Linnaeus cited this figure under A. pumila var. margaritifera in Species Plantarum (p322) as “Comm.hort.2,p.19,t.10.”  Consequently, var. margaritifera and the autonym are synonymous.  Var. pumila, as the autonym, has priority.”

There is more to the reply, but it does not address the problem of the different usages of the name “pumila” by various authors until its inclusion in Haworthia by Duval for a species based on a different type.

There is no confusion in this. It is simply a question of how the Code directs that the name should be treated. I was confused over the issue. I did wade through the detail of synonymy with Mr Larry Leach and subsequently with Dr Peter Bruyns. It was evident from this arduous process that, was “pumila” not available for the species in question, the name “maxima” should be used. There is absolutely no reason for confusion about which species I am referring to, nor about the names available or the way in which they have been used since Scott 1978. But I decided to treat the name “pumila” according to Dr Barrie’s response. This suggested to me that Dr Codd and Col Scott were correct and that “pumila” as used by Aiton and Duval was incorrect and did not preclude the use of the name as typified by Wijnands. The essence is that Linnaeus used the name “pumila” as a prime name in Haworthia and it should sensibly remain there.

What is confusing and what should confuse everyone, is that there are persons who feel the need to contest the issue. The need only arises from personal feelings and the fact that the Code has generated this vast arena for endless vain debate. Where a name has so convincingly and obviously been used in the literature of the time, it obviously meets the need. Changing it, or attempting to change it, generates confusion. We do not have to feed on it.

Haworthia minima and pumila flowers

6645.1b H. pumila

This flower (H. pumila) is apparently persistently regarded by botanists as actinomorphic (star-shaped, radially symmetrical) – as though zygomorphy (yoke shaped, bilateral, asymmetrical) in the aloids is an uncommon condition!

Radial symmetry means the flower can be divided into 3 or more identical sectors which are related to each other by rotation about the centre of the flower. Typically, each sector might contain one tepal or one petal and one sepal and so on. It may or may not be possible to divide the flower into symmetrical halves by the same number of longitudinal planes passing through the axis. Zygomorthic flowers can be divided by only a single plane into two mirror-image halves, much like a yoke or a person’s face.

If you see the way the inner upper petal overlaps BOTH the two lower inner petals, you recognise that there can not be actinomorphy in aloid flowers.

Haworthia pumila

Haworthia minima

MBB7989 Haworthia pumila, Lemoenpoort

Kobus drew my attention to a glabrous plant of H. pumila while he was photographing this species at Lemoenpoort. The plants here have a missing chromosome and tend to have a purplish colour. I have seen smooth non-tubercled leaves elsewhere. But do check out that one plant – if you look carefully you can see the leaves are in 8 nearly vertical tiers. Proper botanists recently, for Taxon, described the arrangement of leaves like this in only two tiers (e.g. H. truncata) as “distichous insertion”. This is weird. Are the leaves in this plant of H. pumila “octichous”? Is H. viscosa “tristichous”. No. The leaves in the aloids are alternately and spirally inserted.

Flower profiles

Flower faces

The absurdity of taxonomy and nomenclature?

In Alsterworthia 13.1:6 (2013) there is yet another statement about the correct name for a species of Haworthia.  It reads …”The vexing matter of the correct name for Haworthia pumila has taxed some of the finest minds in botanical nomenclature”.  The article then goes on to replace that name with H. margaritifera with an explanation so simple that it casts considerable doubt onto the quality of those minds that have examined the problem. There is of course also a difficulty in that the quoted sentence implies that H. pumila is the correct name, while the article goes on to dismiss it.

The fact is that Linnaeus listed four different things (varieties) under a single name Aloe pumila, and the only issue was about which of those four things ends up with that particular first name. It so happened that Burman in 1701 made a choice, and was followed by Aiton in 1789 who chose something else.  So the name Aloe pumila stood but applied to two different things (species). Duval was unaware of the earlier Burman choice and used Aiton’s choice when he created the genus Haworthia.  Dr L. A Codd, whom I would have accepted as a fine mind (and also as a very ethical man), advised C.L. Scott that Aiton’s choice was in fact illegitimate and hence also Duval’s usage in Haworthia.  The opinion was that Burman’s choice was the first and also thus the legitimate one; and it could not be denied by the illegitimacy of the Aiton usage as a later action.  So perhaps it was/is a question of rank or just an opinion that Aiton made up pumila as an entirely new name.

There is nothing complicated and mind boggling about this simple state of affairs. Or is that so?  What on earth does the ICN as the product of presumably fine minds actually say about this? Does it take 50 years of debate to establish such a simple fact? The situation is further exacerbated in that the finest taxonomic minds are involved in an epic battle to either create a single alooid genus or many lesser genera. It appears that the latter option is winning ground although the war against single-species genera has surely not been abandoned. When the dust finally settles, it will be recognised that the taxon onto which this unfortunate species viz. Haworthia pumila/margaritifera resolves, will be a separate genus, probably Tulista, and then the truly correct name will be Tulista pumila. Or will it?

My personal opinion that however the case may truly be judged, the correct answer is the intent. Scott and Codd came to a workable end point way back in history and it has been my misfortune now to have defended that. I think there is a parallel in the case of Aloe bainesii. Put into use by Reynolds far back in history, it is found that the name barberae had page priority and thus preference. In what interest was the change made? Why does the code have a conservation facility for names? The fun seems to be in the argument rather than in usage.

The argument that I think L.A. Codd would have made is this. There are four varieties covered by the Linnaean epithet pumila. The first effective use of that epithet for one of the four was to the warty t10 of Commelin and that is how the name is formally typified.  To use of that same epithet at any other time for any other of those four Linnaean varieties would be illegitimate.  It is also not in the least certain that the name margaritifera is correctly typified by Wijnands on the same Commelin illustration t10. I aided Wijnands to this conclusion before myself stumbling on the fact that its correct typification would be on a Bradley illustration.

There is a curious twist to the issue and it is somewhat of an oversight that the persons involved never read the introduction to Haworthia Revisited. I explained the problem and also in respect to the correct application of the original name margaritifera to what we know as H. minima. I also cited the name H. pumila with the authors as (L.) Bayer to make it clear that I was not accepting the name with the authors (L.)Duval as Scott cited it. I thought at the time that it was a mistake on Scott’s part. I had also written to Dr Codd specifically about the issue and this is when he explained to me (as a professional taxonomist and fine mind) that the illegitimate use of the name pumila in Haworthia did not prevent the correct usage. It only strikes me now that he probably had advised Scott to the effect that the NAME as Duval had taken it through to Haworthia was correct, even if he had applied it to the wrong species. This whole issue has not been properly and fully aired. To argue that it is a new name seems to me just a piece of intellectual vanity that serves no purpose other than to demonstrate our collective failure to honour the intent of the code – or respect the dismay of interested person.  A last point I make is that people can and will always find topics to disagree on, so this is an important trap to avoid and be mindful of. It is not particularly in the case of nomenclature that this seems to happen. I had no doubt at the time when I made the decision to accept Scott’s usage that no matter what my decision was, cause would be found to change it.  If I had decided on either margaritifera (correctly typified) or maxima (as I. Breuer later did), this would also have been argued as wrong.

[-ed. There seems to be a number of taxonomic changes brewing. Time will tell whether pumila survives.]


1. Haworthia margaritifera/pumila
Dr. John Manning. S.ANBI.

The vexing matter of the correct name for Haworthia pumila has taxed some of the finest minds in botanical nomenclature. Since I do not include myself among their company, I was not in the least surprised to find that I had misrepresented the situation. Thanks to expert input from Roy Mottram and Urs Eggli we can now put the matter to rights.

The issue of the correct name for Haworthia pumila starts with the fact that in his original publication of Aloe pumila, which forms the basis for this species, Linnaeus recognized several varieties, but without explicitly listing the typical variety, thus he did not list Aloe pumila L. var. pumila. Linnaeus’ Aloe pumila was subsequently effectively lectotypified by Burman f. (1701) [and later in enor by Scott (1978)] against the illustration in Commelin’s Horti medici Amstelodamensis, which is also the type of var. margarit!fera. This renders the name margaritifera homotypic with Aloe pumila L. (i.e. they share the same type). As the autonym (i.e. following automatically from the species name) for this species, pumila would normally have statutory priority over margarit(fera BUT, in the interim, the combination Haworthia pumila (Aiton) Haw. (1804) had been published, based on the name Aloe arachnoidea var. pumila Aiton, a quite different species that we know now as H herbacea. The publication of this combination renders Haworthia pumila (L.) Duval (1809) an illegitimate later homonym and thus not available for use in Haworthia. Because the combination Haworthia pumila cannot be used for Aloe pumila L. as a result of its prior usage for some other taxon it must be substituted with the next available valid and legitimate epithet, which is margaritifera. Note, however, that in any genus other than Haworthia the epithet pumila is the correct one to be used for this species.

The formal rendering of this situation is as follows:
Haworthia margaritifera (L.) Haw. (1819). Aloe pumila var. margaritifera L. (1753).  Aloe margaritifera (L.) Burm.f (1768).  Aloe pumila L. (1753). H pumila (L.) Duval. (1809), hom. illegit. non H pumila (Ait.) Haw. (1804). Lectotype, effectively designated by Burman f. in Prodromus florae Capensis: 10 (1768) [Superfluous lecotypification by Scott (1985)]: Illustration in Commelin, Horti medici Amstelodamensis, Pars altent: t.l 0 (1701): Aloe Afric: folio in summitate triangulari margaritifera, flore subviridi.

2. International Code of Nomenclature for algae, fungi, and plants (Melbourne Code)

3. Aloe pumila, Haworthia pumila; what or who is confused?  ISBN: 0-9534004-4-1 Bruce Bayer. Alsterworthia International Special Issue No.3. https://haworthia-updates.haworthia.org/aloe-pumila-haworthia-pumila-what-or-who-is-confused/


4. Commelin, Johannes, Horti medici amstelodamensis rariorum tam Orientalis, vol. 2: t. 10 (1701)
Commelin t10 1701 H. pumila


5. Commelin, Johannes, Horti medici amstelodamensis rariorum tam Orientalis, vol. 2: t. 11 (1701)
Commelin t11 1701 H. pumila


6. Curtis’s Botanical Magazine, vol. 33: t. 1360 (1811) [S.T. Edwards]
ST Edward 1811 Curtis Bot Mag v33 t.13608348


7. Moninckx, J., Moninckx atlas, vol. 3: t. 12 (1682-1709)
http://dpc.uba.uva.nl/cgi/i/image/image-idx?c=botanie;view=entry;cc=botanie;entryid=x-421058064 and


8. History of Succulent Plants,  Bradley, Richard (t30) (1716)
Bradley t30


10. History of Succulent Plants,  Bradley, Richard (t21) (1716)
Bradley t21

Ed. – another …

11. J., Moninckx , Aloe Africana, folio in summitate triangulari / Margaritifera, Flore subviridi. / C: Commelin, Hort: Amst: Part: 2. pag: 19. Wijnands, D.O., The botany of the Commelins, Rotterdam 1983, p.134