Now You See It, Now…Why Do You Need So Many Different Eyes?

by Sara E. Freimuth, Claremont McKenna College [edited by Lars Schmitz as part of BIOL 167 “Sensory Evolution”, an upper division class at the W.M. Keck Science Department. Written for educational purposes only].

Imagine you have more than just the two eyes on your face – you have eyes on the top of your head, eyes spanning down your arms and resting on each fingertip, an eye or two at your bellybutton, and even eyes on the bottoms of your feet. Now, imagine that with all of these eyes and all that you can see, you are unable to move beyond choosing to sit or stand in the spot you are right now. You’re probably thinking, if I can’t go anywhere, why do I need all these eyes?

What you are imagining is close to what a fan worm – a sessile marine annelid – looks like and experiences throughout its entire lifetime, and what you are probably thinking is what researchers have been wondering for decades: Why do fan worms have so many complex and diverse radiolar eyes if they live such a sedentary life?

With this question in mind, the radiolar eyes of fan worms, or sabellids, make them an excellent case for examining the emergence of novel visual systems, the development of rudimentary visually guided behaviors, and the function of distributed sensory systems (Bok, Capa, and Nilsson, 2016). The diverse and apparently recently evolved array of eyes in fan worms also surprisingly reveal similarities to both arthropods and vertebrates. Thus, while much has been discovered about the anatomy and physiology of the sabellid visual system, there is still a lot to uncover about what has necessitated its complexity and how it came to be.

fig 1 final

Figure 1. Sabellastarte magnifica, commonly referred to as a feather duster worm, displays its radiolar crown from its tube (Wikimedia Commons).

Fan worms are sessile, tube-dwelling organisms that possess colorful feather-like appendages called radiolar or branchial crowns (Figure 1). They build their protective tubes with mucous secretions, use their radiolar crowns for feeding and respiration, and have a series of segmental bristles that act as a system to help move the fan worms within their tubes and along the sea floor, should they need to abandon their tubes under the most extreme duress.

Despite remaining stationary in their tubes for practically their entire lives, sabellids have a uniquely diverse set of eyes distributed across their entire bodies, and their radiolar eyes, in particular, relate them to other unexpected organisms. While these radiolar eyes, on the one hand, resemble the compound eyes of arthropods in location and function, physiological examination of sabellid radiolar eyes has interestingly related them more closely to vertebrates (Figure 2).

fig 2 final

Figure 2. A summary of the currently accepted molecular phylogeny of the “Big Nine” animal phyla that highlights in red the phylum of the fan worm, Annelida, and the phyla with similar visual structures: Arthropoda and subphylum Vertebrata within Chordata. [constructed as summary of Nielsen, 1995]

Michael Bok, María Capa, and Dan-Eric Nilsson explored the physiology of sabellid radiolar eyes through microscopy, and the combination of these results and the review of more than 100 years of investigations into the anatomical diversity and function of sabellid eyes has substantially improved the understanding of sabellid vision in terms of developmental, evolutionary, and functional significance. The culmination of this research, then, has resulted in the organization and classification of sabellid visual structures into more distinct hierarchies with important evolutionary implications. This visual hierarchy of sabellids yields a perplexing complexity.

At the simplest hierarchical level in the classification of any vision are photoreceptors, which are ambient-light sensitive cells and the primary unit of visual systems. Photoreceptors, historically, have been classified by their differing signaling pathways and structures into two distinct realms: rhabdomeric photoreceptors of invertebrates, such as arthropods, molluscs, and annelids, and ciliary photoreceptors of vertebrates (Arendt, 2003).

fig 3 final

Figure 3. The photoreceptor A shows the general shape and structure of a rhabdomeric photoreceptor found typically in invertebrates, and the photoreceptor B shows that of a ciliary photoreceptor found typically in vertebrates and sabellids. [Figure 3 from Arendt, 2003]

The photoreceptors of sabellids that compose the elements of their visual system within the protective tube are observed to be rhabdomeric in nature. However, results of physiological investigations of the outer, unprotected eyes – the radiolar eyes – of these annelids classify these photoreceptors as ciliary, contradicting the previously established dichotomy and relating the radiolar parts of the visual system of fan worms more closely to vertebrates than to arthropods or other annelids (Bok, Capa, and Nilsson, 2016).

Next up are the ocelli, basic one-pixel light detectors composed of photoreceptive cells that generate directional light sensitivity but not quite full vision. Sabellids have four main types of ocelli – cerebral, segmental, pygidial, and radiolar – that all exhibit an array of visual capabilities for a variety of functions (Figure 4). The former three are all rhabdomeric and housed within the tube of the fan worm, while the latter are ciliary and in multiple types that ultimately compose radiolar eyes.

fig 4 final

Figure 4. A stylized drawing of a sabellid, featuring the various types of ocelli and other notable anatomical features. Radiolar ocelli are unlabeled in this diagram but are included within the radiolar eyes. [Figure 1 from Bok, Cara, and Nilsson, 2016]

The cerebral ocelli, which are buried in the head of the fan worm, typically only have a single rhabdomeric photoreceptive cell, which is reminiscent of arthropods (Bok, Capa, and Nilsson, 2016). Given their location, these cerebral ocelli are useful for little more than metering of ambient light levels. This capability suggests a regulatory function for biological rhythms.

The segmental ocelli also are located within the fan worm tube and consist of two to three rhabdomeric photoreceptors whose illumination invokes a light avoidance response without cerebral input, which suggests these ocelli function to alert the worm that its trunk is exposed.

Even further down in the tube are the pygidial ocelli, which also are of rudimentary rhabdomeric nature and are potentially involved in the quest for a dark location to form a new tube by fan worms that have abandoned their old ones.

The final type of ocelli – radiolar ocelli – populate the radioles of sabellids and consist of four different subtypes classified by their number of cells (Figure 5). Initially, transmission emission microscopy (TEM) found only three types of radiolar ocelli with one, two, and three cells, but Bok, Capa, and Nilsson have used TEM to categorize a fourth type of ocellus with four cells. These researchers even suggest that further investigation of radiolar photoreceptor pigments could yield radiolar ocelli outside these four established realms (Bok, Capa, and Nilsson, 2016).

fig 5 final

Figure 5. Stained thin section light micrographs of the four different radiolar ocelli types in sabelllids (A-D) and their corresponding stylized reconstructions based off these thin sections and transmission electron microscopy studies. The reconstructions are labeled as follows: Nucleus (*), lens (L), sensory cell (Sc), pigment cell (Pc), nervous projections (NP), mitochondrial rods (MR), tapetum (T). [Figure 4 from Bok, Cara, and Nilsson, 2016]

These different types of radiolar ocelli constitute the various forms of radiolar eyes in sabellids, which are part of the next level in the hierarchy of their visual system. Eyes, more generally speaking, consist of at least a single ocellus elaborated with many photoreceptors or cluster together many ocelli in a localized, ordered array. If elaborate enough, eyes have the potential for true image-forming vision in which different photoreceptors sample different points in space.

The radiolar eyes of sabellids can be classified in four different types of eyes: Type S, Type CP, Type CS, and Type CI (Figure 6). While most radiolar eyes fit easily into these ranks, some do not conform clearly to a single classification, and it is important to note that many species of sabellids possess no obvious radiolar eyes at all (Bok, Capa, and Nilsson, 2016).

fig 6 final

Figure 6. Photographic examples of each sabellid radiolar eye type (B) and stylized cross sectional diagrams of each (A). [Figure 3 from Bok, Cara, and Nilsson, 2016]

Type S eyes are composed of scattered single ocelli and are found on both sides of the outside of radioles. They typically occur on every radiole and are occasionally denser on more dorsal radioles. They are the most variable of radiolar eye types in complexity and thus have three subtypes categorized Sa, Sb, and Sc based on the scattering pattern of the ocelli.

Type CP eyes are compound-paired eyes found in several pairs in multiple positions along the outer margins of each radiole. They vary in their presence and pigmentation among different sabellid species and generally decrease in size and number of facets as they approach the tip of the radiole.

Type CS, or compound-single eyes, are similar in that they, too, lie along the outer margins of the radioles, with several inhabiting each. Their development, however, is unique in that these eyes seem to derive from one side of the radiole and grow towards the center, rather than developing in the center and growing towards the sides.

Lastly, Type CI eyes, or compound-inside eyes, are located on the inner margins of the radioles at the distal tip, finishing off whatever chain of eyes the radiole contains. Type CI eyes, like Type S eyes, also vary greatly but in terms of size and location (Bok, Capa, and Nilsson, 2016).

Further investigation and review of each of these eye types, however, yielded no explanations for new or other functions. The current understanding, ultimately, remains unchanged: fan worms use all of these types of radiolar eyes together for only one purpose – to serve as a sort of burglary alert system that triggers a quick withdrawal response where the sabellid retracts its branchial crown back into its tube to avoid consumption by predators (Nilsson, 1994; video).

The single purpose of sabellid eyes for this protective withdrawal response, then, raises a variety of questions. First, this singular purpose potentially suggests that the array of radiolar eyes is not necessarily providing true spatial vision in that the withdrawal response it solicits relies on shadow detection and, thus, may not need to convey any type of image formation by the nervous system. If this is the case, then these eyes only have directional photoreception, rather than true image-forming vision.

A possible explanation to counter this and support true vision is that the neural processing of radiolar visual input in sabellids is more typical of vertebrates than that of other annelids.  There are two types of evidence that may support this explanation. First, one radiole drifting into another’s field of vision does not trigger the withdrawal response, which suggests the radiolar eyes can see some level of depth and supports that they have more complex neural pathways and vision that is more complex than just directional photoreception. Additionally, fan worms have the ability to regenerate their branchial crowns if they are partially removed, which also seems indicative of a more complex neural network, or at least one that is easily altered and extremely resilient (Bok, Capa, and Nilsson, 2016). Both these factors suggest unique neural pathways for sabellids which may yield fascinating results with further investigation.

Another question is why sabellids don’t just use two giant compound eyes to trigger this response instead of the spread of so many different little ones. While this more elegant model of two centralized, economical eyes with great vantage points at the tips of radioles has been successful in other organisms, like arthropods, the dispersal of radiolar eyes seems to account for their vulnerability in the sessile sabellid and may have other potential unknown benefits to be discovered with future exploration (Bok, Capa, and Nilsson, 2016).

The final question that remains is how these eyes evolved. Sabellids have been grouped monophyletically by overall morphology pretty conclusively, but the diversity of their eyes represents anything but a single evolutionary progression (Bok, Capa, and Nilsson, 2016). Interestingly enough, the most complex radiolar eye types of fan worms are not formed by the most complex forms of radiolar ocelli either. With this in mind, it is proposed that sabellids are still contained within one monophyletic group, but there were perhaps at least two or three independent emergences of the radiolar ocelli that constitute their eyes, which likely have just been lost in certain subsequent eye-less genera (Bok, Capa, and Nilsson, 2016).

Moreover, the similarities of sabellid visual structures to both arthropods and vertebrates also remain an evolutionary perplexity that could be better understood with a variety of further investigations. Not only could the aforementioned neural analysis answer questions of both evolutionary and functional similarity of radiolar eyes to arthropods and vertebrates, but an investigation of their opsins, proteins that constitute visual pigments which have been classified in groups similar to those formerly differentiated by rhabdomeric and ciliary photoreceptors, could also clarify (or further complicate) things.

Ultimately, the diverse collection of eyes in fan worms and variety of radiolar eyes in particular posit a unique case of a dispersed sensory system. While the current understanding explores these similarities in functional, physiological, and evolutionary terms, further genetic, developmental, and neurological investigation of sabellid radiolar eyes could yield fascinating insights into the evolution of eyes and visually guided behaviors in general (Bok, Capa, and Nilsson, 2016).


Arendt, Detlev. 2003. “Evolution of Eyes and Photoreceptor Cell Types.” The International Journal of Developmental Biology 47 (7-8): 563–71.

Bok, Michael J., Capa, María, and Nilsson, Dan-Eric. 2016. “Here, There and Everywhere: The Radiolar Eyes of Fan Worms (Annelida, Sabellidae).” Integrative and Comparative Biology 56 (5): 784–95. doi:10.1093/icb/icw089.

Nielsen, Claus. 2012. Animal Evolution: Interrelationships of the Living Phyla. 3rd ed. Oxford: Oxford University Press.

Nilsson, D.-E. 1994. “Eyes as Optical Alarm Systems in Fan Worms and Ark Clams.” Philosophical Transactions of the Royal Society B: Biological Sciences 346 (1316): 195–212. doi:10.1098/rstb.1994.0141.

This entry was posted in sensory evolution blogs, vision and tagged , , , , , . Bookmark the permalink.