contrast between Humboldt’s and de Candolle’s attempt is striking. Whereas Humboldt proposes to use plant form, de Candolle opts to use:
1) Temperature, as determined by distance from the equator, height above sea level and southern or northerly exposure.
2) The mode of watering, which is more or less the quantity of water that reaches the plant, the manner by which water is filtered through the soil and the matter that is dissolved in the water, which may or may not be harmful to the growth of the plant.
3) The degree of soil tenacity or mobility (de Candolle, cited in Ebach and Goujet 2006, p. 768).
Figure 1.3. Carte Botanique de France, pour la 3e Edition de la Flore française par A.G. Dezauche fils Ingénieur Hydrogéologue de la Marine an 13 (1805) “Botanical map of France for the 3rd Edition of Flore française by A.G. Dezauche the son, Marine Hydrological Engineer on the 13th year of the Revolution (1805)” (see Ebach and Goujet 2006, Figure 1.1). For a color version of this figure, see www.iste.co.uk/guilbert/biogeography.zip
The method proposed by de Candolle did not catch on, and by 1820 de Candolle had chosen to use plant distributions instead, dividing the world into 20 regions:
(1) Boreal Asia, Europe, and America; (2) Europe south of the boreal region and north of the Mediterranean; (3) Siberia; (4) the Mediterranean area; (5) eastern Europe to the Black and Caspian Seas; (6) India; (7) China, Indochina, and Japan; (8) Australia; (9) south Africa; (10) east Africa; (11) tropical west Africa; (12) Canary Islands; (13) northern United States; (14) northwest coast of North America; (15) the Antilles; (16) Mexico; (17) tropical America; (18) Chile; (19) southern Brazil and Argentina; (20) Tierra del Fuego (de Candolle, cited in Nelson 1978, pp. 283–284).
Again, de Candolle’s plant regions failed to find acceptance. Humboldtian Joakim Frederik Schouw dismissed the regions:
Candolle compares 20 floras, or as he calls them, regions. In his method, which he has developed studying these floras, [Candolle] does not reveal the characteristics that each form takes; it appears that the main basis for the division [of the regions] is current distributions (Schouw 1823, p. 504, my translation).
So too did his son Alphonse de Candolle, who considered “artificial systems”, which are a detriment to science “when they are considered to be natural” (de Candolle 1855, pp. 1304–1305). So what then is a natural region?
The Humboldtians believed that both biotic and abiotic factors, such as climate, were vital in recognizing plant forms and plant regions:
To have an exact acquaintance with these principal forms of vegetation is of the greatest importance to a phyto-geographical division of the globe, as they principally fix the natural physiognomy of different countries. Humboldt is the first who has made such a classification of vegetation, and this must be taken as the foundation of all further inquiry into the subject. It is not until we are somewhat intimately acquainted with the various characteristic forms of plants, that we will be able to recognise the peculiarities of each flora, and to characterise the physiognomy of each country (Meyen 1846, p. 106).
The problem was deciding which abiotic factors were crucial. Schouw had a list of factors that he thought were crucial, so did Heer (1835), Meyen (1846), Sendtner (1854), Lorenz (1863), von Marilaun (1863) and Grisebach (1872). Schouw’s regions all corresponded to climatic zones (e.g. flora Aplino-arctica) and had a dominant vegetation type (e.g. provincia Cichoriacearum), a practice that the Humboldtians used to define areas. Natural areas (and area classifications) were based on the distributions of vegetation that are driven by climate over time, not dissimilar to the concept of biomes of today. Yet by the end of the 19th century, there was a variety of different area classifications that led Charles Edward Moss to observe that “the subject of ecological plant geography has suffered and still suffers very considerably from a lack of uniformity in the use of its principal terms” (Moss 1910, p. 18).
One idea mentioned by de Candolle (1820) was adopted by the Humboldtians, namely, that of stations and habitations (see Nelson 1978):
By the term station I mean the special nature of the locality in which each species customarily grows; and by the term habitation, a general indication of the country wherein the plant is native. The term station relates essentially to climate, to the terrain of a given place; the term habitation relates to geographical, and even geological, circumstances … The study of stations is, so to speak, botanical topography; the study of habitations, botanical geography … The confusion of these two classes of ideas is one of the causes that have most retarded the science, and that have prevented it from acquiring exactitude” (de Candolle 1820, p. 383, translated in Nelson 1978, p. 280)3.
Stations [habitats] and habitations [regions] were the only natural hierarchy that the Humboldtians adopted. The stations were based on physiognomy and the habitations on endemism. Schouw proposed that “at least half the known species are particular [endemic] to a region; 2. that 1/4 of the genera are either fully [endemic] or mostly occur in a region. 3. that single families are either fully [endemic] or mostly occur in a region (Schouw 1823, p. 504, my translation)”. These rules were not that dissimilar to those proposed by de Candolle: “(1) every species tends to occupy a certain space, and the determination of the laws that govern species distributions is the study of habitations; as (2) there are more species in the tropics than in high latitudes; (3) the numbers of species of monocots and dicots vary in certain ways; (4) certain numbers of species are recorded for certain countries” (de Candolle 1820, pp. 392–400 in Nelson 1978, p. 281). In this sense, you preserve both classifications: plant taxonomy is used to identify plants and species; genera and families are used to determine the larger habitations; and plant forms and vegetation are used to classify the smaller stations. In The History of Biology, Nordenskiöld (1936) summed up plant geography as having taken “two courses”, “a systematical, which is ultimately based on Linnæus’s observations and theories in connexion with the distribution of the plant species, and a morphological [physiognomic], which has its origins in Humboldt’s theories on the morphological association of different vegetable types with different countries and forms of landscape”. Of the latter, Nordenskiöd states, “œcological plant geography does not investigate the nature of the flora, but of the vegetation. It works not with species, but with plant communities [populations]” (Nordenskiöld 1936, pp. 560–561). I will return to the plight of plant geography at the start of the 20th century later.
1.2.4. Zoogeography: a search for natural regions
Animal geography had a later start than plant geography. Although Zimmermann (1778–1783) was the first to consider an animal geography, it was confined to quadrupeds. Unlike the Humboldtians and de Candolle, animal geographers rarely looked at faunal regions, instead preferring to look at taxon-specific distributions. Also, a contemporary of Zimmermann, Johan Christian Fabricius, proposed eight climatic regions “from which the Stations of insects are judged” (Fabricius 1778, p. 154). Zoologists did not adopt the Humboldtian tradition of using “form” and dismissed the climatic regions of Fabricius as arbitrary or artificial:
This simple statement is enough to convince us that there is a lot of arbitrariness in these divisions (Latreille 1815, pp. 40–41, my translation).
[Fabricius] … by not attempting to demonstrate the correctness of any one of his divisions, seems to have subsequently abandoned them altogether, since no one, it may be fairly presumed, was more qualified than himself to discover the artificial nature of his theory (Swainson 1835, pp. 10–11).
Similarly, the regions proposed by Pierre André Latreille in 1817, based on latitudinal and longitudinal gradients along climatic zones, were equally dismissed:
Any division of the globe into climates, by means of equivalent parallels and meridians, wears the appearance of an artificial and arbitrary system, rather than to one according to nature (Kirby and Spence 1828, p. 487).