Contact Us

Use the form on the right to contact us.

You can edit the text in this area, and change where the contact form on the right submits to, by entering edit mode using the modes on the bottom right. 


Santa Cruz, CA

Functional morphologist and evolutionary biologist

Phylogenetics and evolutionary patterns of cleaning

Diversity of cleaners in five families of fishes. Labrids really take the cake!

Diversity of cleaners in five families of fishes. Labrids really take the cake!

Tree inference and phylogenetic methods form a core component of my research. To date, I have used these tools to investigate the lopsided patterns of diversity that we see among cleaner fishes.

The diversity of cleaning behavior is lopsided in more than one way.

First, with at least 67 species that clean, the Labridae (wrasses, parrotfishes, and the lesser-known weed whitings) show a tremendous diversity of fishes that engage in this behavior. This diversity almost 5 times higher than that of the next highest group: the Gobiidae (14 cleaner species). Why does the Labridae have a disproportionately high diversity of cleaners?

Second, over two-thirds of cleaners show the behavior predominately as juveniles and seem to grow out of it as they enter adulthood. The few species that do clean as adults are small - attaining max lengths less than 15 cm. Is cleaning just for small fishes? And do the patterns we see regarding life history stage imply that adult cleaners are neotenous?

Answering these (and other) questions relies on phylogenies that are well-supported and time-calibrated.


I use Bayesian approaches to estimate topologies, branch lengths, and divergence times among species.

I then use the resulting phylogenies to determine when and how traits evolved.


The maximum clade credibility tree from our Bayesian approach. A single (but very representative) character map is painted on the phylogeny. Colors correspond to categories of cleaning: orange is juvenile cleaning, green is facultative cleaning, purple is obligate cleaning, and gray is non-cleaning.

The maximum clade credibility tree from our Bayesian approach. A single (but very representative) character map is painted on the phylogeny. Colors correspond to categories of cleaning: orange is juvenile cleaning, green is facultative cleaning, purple is obligate cleaning, and gray is non-cleaning.

The exceptional diversity of cleaners among wrasses

Because the species richness of cleaners in the Labridae holds plenty of evolutionary intrigue, I first sought to understand how and when cleaning arose in this group (Baliga and Law, 2016). We used maximum likelihood and Bayesian methods to infer the relationships between 320 species of labrids. Using stochastic character mapping, we then simulated character histories across the trees we had generated.

Cleaning itself involves a variety of behavioral and morphological adaptations (see my other research pages!). Species that clean have the capability to target and acquire ectoparasites off the bodies of their 'clients', which often involves acquiescence from these clients. Therefore, one might expect the evolution of cleaning to have relatively few independent origins.

Instead, we found that cleaning evolved at least 30 times within the Labridae, widely dispersed across the tree.

The earliest evolution of cleaning likely occurred 17-21 million years ago. Generally, cleaning appears to be very young: the vast majority of evolutionary transitions likely occurred within the last 10 million years.


Obligate cleaning (purple) evolved at roughly the same time in the Labridae and Gobiidae but in completely separate parts of the world!

Obligate cleaning (purple) evolved at roughly the same time in the Labridae and Gobiidae but in completely separate parts of the world!

Obligate cleaning evolved in separate regions of the world at strikingly similar times

Not only are the Labridae notable for their diversity of cleaners, but they are one of only two families in which highly-specialized obligate cleaning has evolved.

I recently sought to determine if other reef fish groups show similar evolutionary patterning to that of labrids (Baliga and Mehta, 2019). Through this project, we inferred phylogenies for the Gobiidae (focusing on Western Atlantic gobies) as well as the Pomacanthidae (marine angelfishes).


Our tree inference and character mapping showed that highly-specialized obligate cleaning, found in the Indo-Pacific and the Caribbean, evolved in the Labridae and Gobiidae at strikingly similar times: roughly 8-12 million years ago.


Our analyses also extended our findings about the recency of cleaning. Not only has it evolved within the past ~20 million years (or sooner!) within the Labridae, but similar patterns occur in the Gobiidae, Pomacanthidae, Embiotocidae (surfperches), and Pomacentridae (damselfishes).

Why is cleaning so recent? This is still an active area of research, but I suspect that cleaning became an evolutionarily viable strategy 10 - 20 MYA because of two factors. First, evidence suggests that diversification of fishes increased dramatically in the Miocene (particularly on coral reefs!). More fish species could mean a richer pool of clientele for cleaners. Second, ectoparasite diversity (or abundance) could similarly have increased, but presently we know too little to say for sure. Stay tuned!

Density plots that show when transitions to cleaning occurred within each family. Cleaning is quite young, having evolved within the past ~20 million years or sooner.

Density plots that show when transitions to cleaning occurred within each family. Cleaning is quite young, having evolved within the past ~20 million years or sooner.

 
Transitions to cleaning in four other fish families

Transitions to cleaning in four other fish families


Juvenile cleaning is a key evolutionary state

A conceptual flowcart that illustrates how transitions between states occurred over our reconstructions. The width of each arrow is approximately proportional to how frequent the transition occurred.

A conceptual flowcart that illustrates how transitions between states occurred over our reconstructions. The width of each arrow is approximately proportional to how frequent the transition occurred.

The majority of events involved transitions from non-cleaners to the juvenile cleaner state. Cleaning in the juvenile phase appears to be a crucial state that bridges other forms of cleaning (facultative and obligate). This underscores the importance of examining juvenile morphology and ontogenetic patterns of morphological change in this rich and diverse system.