Communicating Phylogeny: Evolutionary Tree Diagrams in Museums

Monophyletic Groups

In the most basic sense, evolutionary trees are branching diagrams showing common ancestry and the relationships between taxa—with variations on this theme depending on the scientific statements being proposed and have a variety of terminology associated with them (e.g., cladograms, phylograms, etc.). One central idea to consider when trying to think about trees phylogenetically, from an evolutionary science and educational perspective, is monophyly (see Wiley 1979, 1981, 2010).

The concept of monophyly as an organizational framework for studying relatedness forms the foundation of phylogenetic thinking, but is often not reflected in classification systems. Donovan and Wilcox (2004) suggest that links to classification in tree diagrams may support the recognition of biological patterns, and research suggests that teaching classification independent of phylogeny supports the development and persistence of alternative conceptions about animal classification (Brumby 1984; Griffiths and Grant 1985; O’Hara 1992; Trowbridge and Mintzes 1988; Wellman and Gelman 1998; Wiley et al. 1991; Yen et al. 2004). The classification schemes that adults and children are exposed to, and most familiar with—such as that birds belong to their own class, Aves, separate from Reptilia—do not reflect the principle of monophyly. The absence of monophyletic groups as an organizational framework for organisms is thought to be particularly problematic for developing an understanding of evolution (American Association for the Advancement of Science 2001; Catley et al. 2005).

More than half of museum trees make links between tree sections and traditional classification categories, and many textbooks present trees alongside widely varying biological classification systems (Catley and Novick 2008). Furthermore, Sandvik (2007) argued that textbooks often adjust the resolution of cladograms—collapse different parts of the tree—to reflect more familiar Linnaean categories, and so these taxa are overrepresented in the diagrams. Whether the predominance of vertebrates in museum trees reflects a deliberate pruning to focus on more familiar Linnaean groups, popular taxa, or institutional research focus is unknown.

From a genealogical perspective, a meaningful classification would reflect monophyletic groups, and the idea of similarity should be understood through the principle of phylogeny. The mismatch between classification and phylogeny can result in grouping by arbitrary (or at least not in evolutionarily meaningful) ways and leads to confusion about shared derived features and convergent similarities. Presenting a phylogenetic tree in conjunction with classification may help novices make connections between the tree and more familiar ideas and ways of thinking, but how best to convey this when these classifications conflict with the statements of relationships depicted in trees is a challenge.

Tree Iconography

Images can be powerful tools for communicating ideas, but their interpretation and understanding are influenced by context and prior conceptions. Visitors’ experiences and understanding of exhibits are framed within a wider cultural framework—with museums challenging or supporting existing knowledge. For example, a study of human evolution museum exhibits by Scott (2007, 2006) found that information about evolution is obtained from a wide range of sources including TV, films, books, family discussions, and museums, and these conceptions influenced visitors’ interpretation and understanding of these exhibits.

Some authors suggest that many of the icons used in evolutionary diagrams—cones of increasing diversity (i.e., trees with narrow bases and wide tops), upwardly directed trees, and trees with differential resolution (emphasizing some taxonomic groups)—reinforce ideas of evolution as progressive and directional (Gould 1995, 1997; O’Hara 1992). Matuk (2007) and Clark (2001) in their discussions of evolutionary images note that the simplified representations of horse evolution that suggest a straightforward and linear progression, first presented in the early 1900s, persists today. Indeed, horse evolution diagrams that depict anagenesis, ancestor–descendant sequences, with taxa arranged sequentially along a time scale, continue to be used in textbooks (Catley and Novick 2008) and are found in museum exhibits.

Unlike biology textbooks (Catley and Novick 2008), “Tree of Life” depictions—diagrams with a central trunk and a distinct “progressive” branching sequence from “lower” organisms on the bottom to “higher” ones at the top (a.k.a “Great Chain of Being” or scala naturae)—were not found in this sample of museum trees. However, two exhibit diagrams have what might be interpreted as vertical, hierarchical representations of primates with a central core and side branches with prosimians at the bottom and apes at the top (Fig. 6). What significance, if any, visitors might attribute to these particular examples is unknown, but previous work has demonstrated the potential for interpreting the layout of exhibits that include humans, their most recent extinct relatives, and/or other primates as directional and progressive (Scott and Giusti 2006).

Orientation and Direction

While identical evolutionary relationships can be depicted using any tree orientation and direction and/or geometrical shape, the particular form used may have implications for its accessibility to users and impact the interpretation of information shown in the diagram. Spatial framework theory suggests that the directions used to refer to something are based on the participants using their body as a reference point and that biases in our perceptions of horizontal and vertical space result from our conceptual representations of those spaces (Franklin and Tversky 1990; Tversky 2002, 2005a). Cross-cultural studies have found that directionality varies by concept and language but that both children and adults map temporal increases horizontally on diagrams, with the direction of time reflecting the direction of their written language (Tversky 2001, 2005b; Tversky et al. 1991).

The potential implication for orientation of tree diagrams is two-fold: misreading of time direction and the potential for reinforcing linear and progressive conceptions of evolution. The misreading of time across the top from left to right (in vertically oriented trees) rather than from bottom-up is a common misconception in interpreting tree diagrams (Giusti and Scott 2006; Gregory 2008; Meir et al. 2007).

The majority of museum trees sampled were oriented vertically with branches directed upward from the root. It is possible, given perceptual biases of horizontal and vertical space, that the tendency toward using vertical and upwardly directed diagrams contributes to this common error in reading temporal direction on trees. Vertically oriented diagrams have the potential to create confusion about the direction of time, particularly when not all trees explicitly label time, either absolutely or relatively. Many tree of life depictions in biology textbooks have no direct indicator of time, leaving it to the user to determine the relative time direction which may be incorrectly inferred (Catley and Novick 2008). In this study, fewer than half of museum trees label time on the diagram, but many depict time diagrammatically through variation in branch length for extinct and extant taxa.

Also, it is possible that vertical trees have the potential to reinforce ideas of progression and direction in evolution, as vertically oriented diagrams are often associated with quantitative increases and notably correspond to the linguistic metaphors of “up” and their associations with concepts of “more” and “better” (Tversky et al. 1991). The idea that evolution is a directional process from lower/primitive to higher/advanced is a powerful cultural narrative, often mirrored in popular imagery about evolution (Clark 2001; Gould 1997; Green and Shapely 2005; Matuk 2007; O’Hara 1992).

However, Phillips et al. (2010) found that the layout of terminal taxa in a cladogram that is oriented horizontally from root to terminal points—so that the taxa are organized vertically along the edge—elicits more frequent teleological responses and explanations from students than cladograms oriented vertically from the root, where terminal taxa are organized on the horizontal. Therefore, the authors suggest using cladograms with terminal taxa oriented horizontally—a vertical root to branch orientation—and the placement of more complex taxa in the middle to help avoid teleological thinking. These results support the embodied cognition perspective discussed earlier (Franklin and Tversky 1990) but differ in the tree element being considered in the context of orientation—overall tree or resultant layout of terminal taxa.

Furthermore, learning research has found that reasoning about evolution differs by organism (Diamond and Evans 2007) and that the interpretation of cladograms is impacted by users’ prior knowledge, and their narratives about evolution are typically overlain onto tree diagrams (Matuk 2008a, b, c; Matuk and Uttal 2012). The relative importance of overall tree orientation and conceptions of diagrammatic space and the layout of terminal taxa as a result of that orientation—and how either or both may be ameliorated—warrants further consideration.

In addition to orientation, geometry has implications for tree understanding. While different geometries show equivalent relationships and the selection of one versus another may be arbitrary, the particular form used may have implications for interpretation. Novick and Catley (2007) found that undergraduate students had greater difficulties extracting the hierarchical structure and relationships in angled cladograms than rectangular ones (what they refer to as ladders and trees, respectively) despite their being equivalent in terms of the information they contain. The authors suggest that the difficulty in seeing the nested relationships in the ladder results from the Gestalt principle of good continuation. Good continuation implies that the sloped line at the base of the ladder/angled diagram represents a single hierarchical level rather than the multiple levels it actually represents. The principle of good continuation then acts as a cognitive constraint resulting in the straight line being seen as a unit that continues without change, making it difficult for students to understand and interpret the relationships being depicted. Angled cladograms were also found to be more likely to elicit anagenic responses—speciation by transformation of one form into another—than rectangular ones (Novick et al. 2010b).

Humans in Evolutionary Trees

Museum visitors’ reasoning about organisms and evolutionary explanations varies depending on the taxa included in the tree diagram, particularly humans (Diamond and Evans 2007). How visitors perceive exhibits with humans and other living or extinct primates in them is complex and challenging, but they are often interpreted as being linear, directional, and progressive (Scott 2007, 2010; Scott and Giusti 2006).

In addition to the common vertical orientation of trees, the location of Homo sapiens and other hominin species in relation to the other taxa in the tree has the potential to reflect and reinforce ideas of teleology and progression (Matuk 2007; Tversky 1995). A survey of textbook charts found most to be vertically organized with H. sapiens at the top (Tversky 1995), and an analysis of anthropocentrism in phylogenetic textbooks found the position of humans on the top-right of the left–right axis of vertical cladograms to be significant (Sandvik 2007). In museum trees, a bias for top-right placement of humans was not found; however, the sample size was small (n = 9).

The common misreading of time across the top of a cladogram from left to right—coupled with reading the order of terminal taxa across the top as relatedness—may be interpreted as a progression from “old, primitive or simple” to “recent and complex,” culminating in humans (Baum et al. 2005; Catley and Novick 2006; Giusti and Scott 2006; Halverson et al. 2008; Meir et al. 2007). Furthermore, a recent study of the impact of taxa placement in cladograms found that students were more likely to provide teleological responses and explanations if humans occupied an end, rather than a central location (Phillips et al. 2010).

In addition to the placement of H. sapiens, the portrayal of hominin evolution as primarily anagenic, by depicting one or more taxa placed on or within a single branch, is problematic for its potential to reinforce ideas of teleology, progression, and anthropocentrism. While anagenesis is common in textbook trees with humans (Catley and Novick 2008), fewer than a third of museum trees that include humans depict anagenesis. However, of those that do, all include H. sapiens and their most recent extinct relatives (e.g., Homo, Australopithecus, etc.) rather than humans in relation to other extant primates or other taxa.

Geological Time

Time is an important and difficult concept in understanding evolutionary trees, and the interpretation of time on trees is influenced by a range of factors including branch length and naïve understanding of evolutionary processes (Dodick 2010). It has been suggested that where temporal data are available, the inclusion of geological time on diagrams may help to support understanding (Catley and Novick 2008) and help with the common misreading of time across the top rather than bottom-up in vertically oriented trees (Meir et al. 2007).

Variation in branch length is thought to have the potential to promote understanding if the earlier ending points indicate extinct taxa (Catley and Novick 2008), and the inclusion of extinct taxa could help to avoid ideas of species persistence and progress (Donovan and Hornack 2004)—in part because a long branch is often incorrectly interpreted as a lineage in which no change has occurred (Crisp and Cook 2005; Novick and Catley 2007). However, the potential value of different branch length to identify extinct groups may be hampered by the fact that the significance of this diagrammatic feature is often not made explicit. Recent research indicates that there is a strong correlation between understanding the direction of time and the ability to explain evolutionary problems as represented in phylogenetic diagrams and that explicitly including temporal information on diagrams may support understanding and avoid the common misreading of relatedness along the tips, a.k.a tip-reading (Dodick 2010). The interpretation of time in phylogenetic trees and advantages and disadvantages of explicitly doing so are subject to much discussion, and continuing research will help to clarify these issues (Catley and Novick 2008; Dodick 2010).

Tree Content and Labeling

Most of the museum trees do not label tree components such as the root or node(s) as representing common ancestors or shared derived characters (synapomorphies) between taxa. Fewer than half refer to ancestors/common ancestors—with less than 15% included on the diagram itself—and only 20% labeling specific synapomorphies that support the relationships depicted on the tree. Donovan and Wilcox (2004) suggest that labeling the root or other internal node as “common ancestor” can help to overcome the abstractness of tree representations and support the interpretation of nodes. Others argue that since the ancestor is unknown, it is disingenuous to include it (Catley and Novick 2008), and doing so has the potential to reinforce the view of nodes as precise moments of change (Meir et al. 2007). Recent research has found that the inclusion of synapomorphies can help support understanding of tree diagrams and that evolutionary relationships are based on shared characteristics (Novick et al. 2010a), making their relatively uncommon use in museum tree diagrams problematic in terms of supporting visitor understanding.

Museum trees often include, in graphic form, other information beyond relatedness and common ancestry, such as diversity, by altering variables such as branch length, thickness, or shape and using color-coding and symbols. These are often not made explicit on the tree itself or in an associated legend or key. Textbook trees often use branch thickness to indicate diversity, but the graphical significance of this is generally unclear and undefined (Catley and Novick 2008). The absence of clear labeling means that the significance of these variables, if any, may be unclear, which makes it difficult to read and interpret the diagram. Being explicit about the intent of abstract diagrammatic elements is likely to aid in tree interpretation.

Tree Explanatory Information

For most museum trees, the exhibit text describes what can be seen in the tree—e.g., which taxa are most closely related—but the link to the graphic itself is usually not explicit. Evaluation studies suggest that it is important to directly tie labels to what visitors can experience at that point in the exhibition (McLean 1993; Serrell 1996), and presenting explicit information and concrete ideas in exhibit labels helps to instruct visitors about what they should look for (Bitgood 2000; Falk 1997; Falk and Dierking 1992). However, the lack of explicit annotation in many museum trees is not surprising given its absence in most evolutionary diagrams used in textbooks (Catley and Novick 2008), although its inclusion could support an understanding and interpretation of evolutionary processes (Donovan and Hornack 2004). Overall, the absence of explicit explanations for many trees or information about trees as products of science is likely to add to the difficulty that visitors have in reading and understanding of these diagrams.

Tree Presentation

Overwhelmingly, tree diagrams used in museum exhibits are part of graphic panels with images or specimens/models of taxa at terminal taxa points. Incorporating visuals into trees may draw attention to the organisms, help users to recognize and identify taxa, and assist visitors in connecting labeled synapomorphies with visible morphological characteristics. Many novices emphasize morphological features and similarity-based reasoning in their thinking about biological relationships, and so caution should be used to avoid conflating overall similarly with relatedness (Gelman 2004; Gelman and Markman 1987; Halverson et al. 2008; Sloutsky et al. 2001); however, explicitly labeling synapomorphies that are used to support the relationships shown in the tree—and perhaps that can be seen in accompanying visuals—may help highlight the evidence used in tree building, that of relatedness based on shared derived characters, and support ideas about scientific inference (Donovan and Wilcox 2004).

Fewer than 10% are multimedia-based, but some of these kiosks and online trees were interactive, where the user could step through the information or navigate to different parts of the tree. Summative evaluation of Yale’s Travels in the Great Tree of Life exhibit found that the computer game exploring relationships was effective at communicating the idea that phylogenetic relationships may not always be what you might expect (Giusti 2008), which suggests that interactivity and/or animation may help address some issues with reasoning using trees. Based on personal experience with museum visitors, exploring the tree of life using manipulatives such as using scale models of taxa and different graphic representations can be effective with museum visitors. Research on the potential role of animation in understanding cladograms has found that animations can influence the perception and interpretation of diagrams, but that interpretation is also impacted by a user’s prior knowledge and common evolution narratives (Matuk 2008a, b, c, 2010; Matuk and Uttal 2012).