This study is an attempt to expand a previous survey by Fisler and Lecointre (FL) for systematizing ideas on the use of the tree metaphor in classification, as expressed by various historically important figures in their writings. FL used a cladistic approach to analyze their data, as employed in biological classification. We supplement this analysis here using several methods of multivariate data exploration, producing a UPGMA dendrogram, a minimum spanning tree, a neighbor joining additive tree, a plexus graph, a phylogenetic network, and two multidimensional scaling ordinations of the same data used by FL. We confirm the validity of many of FL's smaller clusters of writings, and revealed a new 3-group categorization undetected by the previous study. These three groups largely correspond to Classifiers, who did not consider evolution for historical reasons or on purpose, Non-analytical evolutionists, who recognized evolution but with a more or less naive attitude towards the temporal change of life, and Modelers, with more explicit views on evolutionary processes, often applying objective mathematical tools for exploring the past and present of organismal diversity. Some scientists were difficult to assign to any group unambiguously, including J.W. von Goethe, who takes a unique position in the history of biology, and, to a lesser extent, E. Mayr and G.G. Simpson, the leaders of the gradist school of systematics. We argue that cladistic methods are insufficient by themselves, notably in situations where there are no obvious ancestor-descendant relationships underlying the development of the objects being analyzed.