Why is this? Shortly after Humboldt and Bonpland returned to Europe in 1804, they split their large collection, with each receiving an unknown number of specimens, likely at least 6000 for each (Stauffer et al., 2012). Bonpland then took his part of the collection to Argentina where he decided to settle in Santa in Ana November 1816. Apparently the division of their material in 1804/1805 did not involve a simple separation of duplicates because as early as 1806, Humboldt begged Bonpland in at least three letters to return specimens belonging to what he considered his set (Stauffer et al., 2012). This never happened, but Bonpland did send part of his herbarium, including thousands of collections from their 15,000 km-joint travels, from Argentina to Paris in 1832. In Paris, this material was integrated into the general herbarium, without a list being made of the contents (Stauffer et al., 2012). Locating these specimens among the 5.7 million specimens in the Paris herbarium (Le Bras et al., 2017) is an outstanding task. After Bonpland’s death, the remainder of his herbarium also came to Paris and was likewise intercalated into the general herbarium.

From his own set, Humboldt gave at least 3525 specimens to the herbarium in Paris. These plants they have been fully digitised and databased (Le Bras et al., 2017). He also sent at least 3360 specimens to his friend Willdenow in Berlin, where they survived World War II because the Willdenow herbarium had been kept as a separate historic collection and been evacuated during the war (Hiepko, 2006; Stauffer et al., 2012). By contrast, Allied bombing destroyed other Humboldt specimens deposited in the general herbarium in Berlin, and they only survive in the form of Black/White photos made by the American taxonomist J. F. Macbride shortly before the war (Macbride photos of the Willdenow herbarium in Berlin).

For these reasons – and further complications described by Stauffer et al. (2012) – we do not know the full number of plant collections made by Bonpland and Humboldt, the current whereabouts of all of these specimens, or, as a consequence, their precise collecting locations. This, and the fact that Bonpland and Humboldt did not number all their collections (but only some), greatly handicaps the use of this precious material for ecological and climate-change research. For example, Humboldt’s famous 1807 Tableau Physique of Mt. Chimborazo (https://www.avhumboldt.de) and accompanying Essay on the Geography of Plants lists plants collected in other parts of Ecuador (mostly on Mt. Antisana) and even South America, but this is difficult to verify (Morueta-Holme et al., 2015, 2016; Sklenár, 2015; Moret et al., 2019). A clear example is Espeletia, which Humboldt’s drawing shows as occurring high on the Chimborazo, but which has never been recorded from that mountain (Sklenár, 2015). Uncertainty regarding collecting sites has obvious implications for the altitude at which species occur or occurred in the past.

Using image-recognition tools to analyse the labels of digitized herbarium specimens could help find all relevant material in all herbaria with Bonpland and Humboldt material (Stauffer et al., 2012), and the Paris specimens could provide a kind of backbone for such a project since the full content of that herbarium has been digitally imaged (Le Bras et al., 2017). Adding the Berlin-Willdenow herbarium, which is also available online, and other relevant herbaria, would provide a larger dataset. Analysing this with machine-learning tools  could lead to improved accessibility of Bonpland and Humboldt material as well as better georeferencing of their collecting localities. Today, 215 years after their important expedition, it seems about time we do this!

Acknowledgements: I thank Pierre Moret, Petr Sklenár, Marc Pignal, and Larry Dorr for their comments on this blog.

Hiepko, P. 2006. Humboldt, his botanical mentor Willdenow, and the fate of the collectionsof Humboldt & Bonpland. Bot. Jahrb. Syst. 126: 509–516.

Humboldt, A. von, Bonpland, A. (1807) Essai sur la Géographie des Plantes, Accompagné d’un Tableau Physique des Régions Equinoxiales (Fr. Schoell Libraire, Paris). Available at botanicus.org/title/b12218212. Accessed November 1, 2015. 

Le Bras, G., Pignal, M., Jeanson, M.L., Muller, S., Aupic, C., and 18 others. 2017. The French Muséum national d’histoire naturelle vascular plant herbarium collection dataset. Sci. Data 4:170016 doi: 10.1038/sdata.2017.16.

Moret, P., Muriel, P., Jaramillo, R., Dangles, O. 2019. Humboldt’s Tableau Physique revisited. Proc. Natl. Acad. Sci. USA www.pnas.org/cgi/doi/10.1073/pnas.1904585116

Morueta-Holme, N., Engemann, K., Sandoval-Acuña, P., Jonas, J.D., Segnitz, R.M., Svenning, J.C. 2015. Strong upslope shifts in Chimborazo’s vegetation over two centuries since Humboldt. Proc. Natl. Acad. Sci. USA 112: 12741–12745.

Morueta-Holme, N., Engemann, K., Sandoval-Acuña, P., Jonas, J.D., Segnitz, R.M., Svenning, J.C. 2016. Reply to Sklenar. Proc. Natl. Acad. Sci. USA 113: E409–E410

Sklenár, P. 2015. Advance of plant species on slopes of the Chimborazo volcano (Ecuador) calculated based on unreliable data. Proc. Natl. Acad. Sci. USA 113: E407–E408. 

Stauffer, F.W., Stauffer, J., Dorr, L.J. 2012. Bonpland and Humboldt specimens, field notes, and herbaria; new insights from a study of the monocotyledons collected in Venezuela. Candollea 67: 75–130.