Like many other taxa in section Otidia, Pelargonium hermansdorpense has been difficult to develop a clear understanding of. The taxon was described by Knuth in 1930 [1], however, the Berlin type specimen was lost following the Allied air raid in the night of 1-2 March 1943, when a significant portion of the world’s most important botanical collections at Dahlem was lost in a fire destroying about 4 million specimens. Luckily, the isotype at the Bolus herbarium survived, which now enables us to understand Knuth’s description much more clearly.

Knuth’s description (Fig. 1) highlights succulent stems about 7 mm thick, membraneous laminas ovate 20.0-25.0 x 10.0-12.5 mm, lobed with crenulate margins, peduncles up to 110 mm, inflorescences with 8-10 flowers, pedicels 10-15 mm, nectar tubes 5-6 mm. Petals 3, oblong, stamens 5.

Fig. 1: Knuth’s description of P. hermansdorpense [1].

Knuth places the taxon in the section Otidia and describes the stems as being 7 mm thick, which would correspond to the thinner-stemmed otidias, i.e. not carnosum or most parviflorum. His most unusual claim was that P. hermansdorpense had “only 3 petals”, which must have seemed odd to him, as no other pelargonium has three petals and we must treat this as an error.

Knuth’s other error in the description was the reference to the locality, i.e. Hans Meys Kraal in the “Hermansdorp” district. The farm can easily be located near Humandsorp of course and lies at the eastern end of the Baavianskloof Wilderness Area, close to Andrieskraal and Patensie.

Fig. 2: The type specimen of P. hermansdorpense at the Bolus Herbarium, Cape Town, BOL 150165. Note the shallowly lobed laminas in the blue cut-out and the short petals visible in the yellow cut-out.

The type specimen at the Kirstenbosch Botanical Gardens (Fig. 2) shows two branches bearing branched inflorescences with pseudoumbels of flowers with long pedicels and short nectar tubes. There are several otidias with pedicels much longer than nectar tubes. P. crithmifolium, paniculatum and klinghardtense are clearly unrelated as they have much-branched inflorescences with thick stems. P. adriaanii is a Richtersveld species related to P. carnosum with thick stems also. P. dasyphyllum forms much-branched shrubs with thinner stems than either of the above, however, its petals are long and narrow, much longer than the sepals. This leaves us with P. laxum and P. keeromsbergense, with which P. humansdorpense could be related, given the short petals visible in Fig. 2. (yellow cut-out). However, P. keeromsbergense is a rather large species with thick upright stems, much thicker than those in BOL 150165. By exclusion, we thus come to P. laxum as potentially the closest taxon.

P. hermansdorpense or P. humansdorpense or P. onustifolium?

Much ink has been spilled on this matter, and Victor [2] provides an excellent account of it. Name correction was attempted by Clifford in 1958 [3] who suggested P. humansdorpense to correct the evident orthographic error. Without reference to Clifford, in his review of section Otidia, Vorster [4] also refers to P. humansdorpense when saying that material was found by “Mr. Brian Makin […] in the area between Touwsrivier and Montagu which looks suspiciously like this species.” In his PhD Thesis, Becker in 2006 [5] returns to the correct binomial P. hermansdorpense, and establishes a close relationship between it and P. laxum on the basis of their shared distribution in the Eastern Cape Province and the long pedicel, but does not dwell on the subject matter further. Becker refers to the specimen in Fig. 2 as BOL 107081, however, this has later evidently been re-numbered to BOL 150165.

Vorster’s asserted likeness of Makin’s find to BOL 150165 and use of the incorrect name of “P. humansdorpense” has led many into an unusual rabbit hole. While it hasn’t been possible to re-locate Makin’s population, Powrie and Van Jaarsveld, as well as others, collected plants, featured in Victor’s article [2], which all had thin stems and occasionally shallowly lobed leaves, but had very long linear petals, unlike BOL 150165. Becker described these as P. laxum ssp. karooicum in 2008 [6], however, in 2014, Clifton deposited a plant, descending from Powrie’s collection at Anysberg, via Fibrex Nursery, as “P. humansdorpense” in the Kew herbarium (specimen K000904399). We now know that this is the typical Anysberg P. laxum ssp. karooicum, but it is mind-boggling how one error led to another and now we have “P. humansdorpense”, a taxon that is entirely imaginary, permeating the literature and established herbaria, Kew of all places.

However, the story does not end here. By 2015, Clifton became convinced that the Anysberg taxon deserved to be described as a species of its own, and introduced P. onustifolium [7]. I commented on this taxon in 2018 [8], as well as on the dangers of describing otidias from cultivated material [9], but it bears repeating that there are no other differences between P. laxum ssp. karooicum and Clifton’s P. onustifolium than the lamina shape, which is more shallowly lobed in the latter.

Both develop a central tuber and side tubers, which is a character Clifton didn’t notice, possibly because his plants were grown from a cutting. P. onustifolium should thus be treated as a synonym for P. laxum ssp. karooicum. The view is supported by Becker’s analysis as well [6].

Hans Meys Kraal

Hans Meys Kraal, where P. humansdorpense was collected by Fourcade in 1927, is a farm crossed by the Kouga River just before its confluence with the Groot River, forming the Gamtoos River, flowing past the nearby settlements of Andrieskraal and Patensie. For those interested, Victor [3] provides an interesting summary of Fourcade’s activity in South Africa, which led to the description of further two Pelargonium taxa, i.e. P. gracillimum and P. molicomum.

The location is in a geologically incredibly interesting area running along the Gamtoos fault at the border between Table Mountain and pre-Cape strata [7], which represent the basis that was folded and eroded before the sedimentary sequences of the Cape Supergroup (such as the Table Mountain Group) were deposited over them. It would not be unusual to expect unusual species to develop in such areas, such as the incredibly localised P. campestre, found in the flats around Gqeberha (formerly Port Elizabeth).

Fig. 3: Jean-Andre Audissou’s observation from Patensie, 2010. Reproduced with permission.

Victor writes in 2014 [11] that “through my friend Cameron McMaster, I arranged for a group of six people from Cape Nature” to visit the original location. “They combed the area over the course of a weekend, but could not find any trace” of P. hermansdorpense. While the farm is now part of the Baviaankloof Wilderness Area, the areas close to the Kouga River are heavily developed.

Thus, to find the taxon it would be meaningful to look for similar habitats at about 150 m a.s.l. along the Kouga, Groot and Gamtoos rivers.

 

Recent, and not-so-recent observations

It recently transpired that the renowned French plantsman Jean-Andre Audissou visited the area already in 2010, and found plants that resemble BOL 150165 in almost all characters (Fig. 3). The leaves are slightly more deeply lobed, and the stems look thicker than Knuth’s 7 mm, but it would appear that the photos depict plants that match the herbarium specimen pretty closely. Unfortunately, Audissou does not recall the exact location except that it was close to Patensie, which is about 20 km to the East of Hans Meys Kraal as the crow flies.

There are other more recent observations, e.g. iNat 106972274 from 2022 (Fig. 4), from Andrieskraal [12], which is about 10 km to the East and iNat 218848901 from 2024, from close to Cambria [13], about 8 km to the North of Hans Meys Kraal. Common to these observations is that the petals are about as long as the sepals, and barely marked, which is consistent with P. laxum ssp. laxum. The length is also consistent with BOL 150165, although it is not possible to see if the petals of the latter are marked and it is no surprise that there are 5 of them in Fig. 3.

The leaves of the three recent observations of P. laxum ssp. laxum from this area are all variously lobed, from deeply lobed in Fig. 3, to only deeply crenulate in Fig. 4, all of which is consistent with BOL 150165.

This lamina shape is less consistent with the description of P. laxum ssp. laxum, which is provided by Becker [6] as “bipinnate and toothed”. However, while many inland populations of P. laxum ssp. laxum have leaves that correspond to Becker’s description, e.g. those around Graaf Reinet and Cradock, plants elsewhere exhibit variously deeply and broadly lobed and pinnate or bipinnate laminas. It would thus appear that the leaf shape in P. laxum ssp. laxum is quite variable and could easily accommodate the shallowly lobed leaves of BOL 150165.

A character of BOL 150165 that is not evident in the recent observations by Audissou, Molteno and Hankey, is the reduced stem width. The dried stems measure about 5 mm in diameter, which is consistent with shrinkage expected during drying if the live stems were 7 mm wide. It is difficult to gage the width of the stems in Figs. 3 and 4, however, if compared to the leaf widths, the narrower branches in Fig. 4 could be about 1 cm wide, which is very close to Knuth’s description.

Fig. 4: Steven Molteno’s observation from Andrieskraal, 2022, iNat 106972274 [12]. Reproduced with permission.

Conclusion

Through iNat and social media, citizen science has brought to light several observations of populations that closely resemble BOL 150165, on the basis of which Knuth’s P. hermansdorpense was described. The isotype shows two branches with inflorescences, the structure of which is consistent with P. laxum ssp. laxum. Although the petals in BOL 150165 are difficult to discern, they also seem to be consistent with laxum. The leaves are much less shallowly lobed than in laxum, however, this taxon shows a significant variability in leaf shape, within which the leaves of BOL 150165 present less deeply incised margins. The stems of BOL 150165 are also at the thin end of what would be considered usual for laxum (1-2 cm), however, while the 7-mm diameter observed in BOL 150165 appears less than the branches of the observations by Audissou, Molteno and Hankey, some of the side branches in the latter do look much thinner than 2 cm.

Fig. 5: The root systems of P. laxum ssp. laxum near Graaf Reinet, showing thickened taproots with thin side roots. This diverges from Becker’s description of the root system of laxum ssp. laxum as “simple”. The plants also present branches that are thinner than 1 cm, demonstrating the great variety of stem thicknesses in P. laxum ssp. laxum. The plants were replanted after the photograph was taken.

The root system in BOL 150165 is not mentioned by Knuth, unfortunately, and neither is it presented in the herbarium specimen. Becker [6] describes P. laxum ssp. laxum as having a “simple root system”, however, we know that it develops a central taproot, as is clearly seen in Fig. 5, that can extend to 20 cm or more into the ground in older plants.

While there are some interesting peculiarities of BOL 150165, it appears that it largely falls within the description of P. laxum ssp. laxum, although studies on live specimens would be useful to ascertain the characters that are difficult to observe in photographs, such as typology of hairs on leaves and petals. Otidia laxa was described already in 1824 by Sweet [14] and it is clearly the name that takes precedence, meaning that P. hermansdorpense should be treated as a synonym of it.

Matija Strlič, Ljubljana, Slovenia

 

Literature

1. Knuth (1930): Pelargonium hermansdorpense, Repert. Spec. Nov. Regni Veg. 28, 90.
2. Victor (2012): A little known species of Pelargonium Section Otidia (Geraniaceae): Pelargonium hermansdorpense including the confirmation of its correct name, Geraniaceae Group Assoc. Notes no. 66.
3. Clifford (1958): Pelargoniums including the popular ‘Geranium’, Blandford Press, Chatham, pp. 195, 199.
4. J. Vorster (1990): Taxonomy of the genus Pelargonium: Review of the section Otidia. In: Proceedings of the International Geraniaceae Symposium, University of Stellenbosch, 24-26 September 1990, reprinted by the Geraniaceae Group, 1994, pp. 279-294.
5. Becker (2006), Revision der Pelargonium – Sektion Otidia (Geraniaceae) aus dem Winterregengebiet des südlichen Afrikas und Bewertung evolutiver Strategien der Pelargonien aus der Capensis. PhD Thesis, University of Münster, p. 108.
6. Becker, K. Schäper, F. Albers (2008): Description of two new taxa of Pelargonium section Otidia (Geraniaceae), P. keeromsbergense and P. laxum ssp. karooicum. Schumannia 5, 181-190.
7. Clifton R. (2015), Pelargonium section Otidia: diagnosis of the new name onustifolium. Geraniaceae Group Associated Notes no. 81, 28-30.
8. Strlič (2018): The Laxum Conundrum. Geraniaceae Group News #150, pp 15-24. Available online: https://pelargonium.si/wp/2018/06/07/note-8-the-laxum-conundrum/ (15/03/2026).
9. Strlič (2016): Otidias, or the Story of Dürer’s Rhino, Pelargonium Notes, #4. Available online: https://pelargonium.si/wp/2016/12/29/note-4-otidias-or-the-story-of-durers-rhino/ (15/03/2026).
10. W. Shone, C. C. Nolte, P. W. K. Booth (1990): Pre-Cape rocks of the Gamtoos area / a complex tectonostratigraphic package preserved as a horst block, S. Afr. J. Geol., 93, 616-621.
11. Victor (2014): Pelargonium hermansdorpense, Geraniaceae Group News no. 132, 25-29.
12. Steven Molteno (2022), https://www.inaturalist.org/observations/106972274 (15/03/2026).
13. Andre Hankey (2024), https://www.inaturalist.org/observations/218848901 (15/03/2026).
14. Sweet (1824): Otidia laxa, in: Geraniaceae: t. 196. (Iconotype reproduced in [8]).

 

Citation and Copyright

© The Author. This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.

ISSN 2464-014X.

This article was first published in Geraniaceae Group News #180. Cite as: M. Strlič (2026): On Pelargonium hermansdorpense. Geraniaceae Group News #180, pp 21-31.