Peer Review/Input on beginning of O. cyanea paper

Joined
Dec 16, 2005
Messages
672
Location
Laie, Hawaii
Aloha everyone,

Several months ago I posted about my undergraduate research project (and asked for ideas for possible other projects), and have decided to carry on with the O. cyanea research as originally planned. I am getting ready to start the actual experiment, and would like some of your professional input and suggestions.

Please read my paper here: Google Drive: Sign-in

Some questions/specific things I am looking for help or advice on are:

1. Would the octopuses do okay in a bare bottom tank? Originally I was thinking of covering 3 sides of a cube tank with the desired background, and leaving an inch or two of sand in the bottom as well as some rocks. However, if the octopuses would not be uncomfortable in a bare-bottom tank, I could also find a way to put the background color underneath. My thinking is that articles I have read (about cuttlefish) indicate that background matching may be more influenced by what is underneath them rather than beside them. Also, I want the octopus to be comfortable so it acts as naturally as possible, but I also don't want him to hide all the time. Would a few medium size rocks they could use as a den be sufficient to keep them happy?

2. What do you think of the time frame for each background? In the paper I have each background for 2 weeks, but, partially due to time I am now thinking about shortening that to one week.

3. Finally, and the most perplexing problem I am facing at the moment: what would be a statistical way to analyze my data? The experiment is based off of behaviour and observations, but how can I measure differences in a scientific way?

If you know of any other papers that aren't in my works cited that could be useful, please let me know. Any other suggestions about how to refine this project are welcome!

Thanks!
Nick
 
I am a bit confused on the duration. Why 1 or 2 weeks and not minutes or hours? The rapid change is what they are most noted for so I am missing something with the extended exposure parameter. It would seem that keeping them in normal reef environment tanks and then transfering them to something unfamiliar and videoing the process might be a way to go unless you are expecting changes over a long period of time (inking might be an issue). Varying amounts of time (say 3 minutes to 8 hours) might also give interesting results.

For quantifying your results, look into Hanlon's 3 camouflage types (uniform, mottled and disruptive) and get a clear idea how each is expressed so that you can identify the characteristics. It may be that unnatural backgrounds will always produce one of the three or may be reliant on how well they can produce the unusual colors using their yellow, red and brown chromatophores and the underlying reflections (green should be a challenge and blue likewise interesting).

You might also use rocks (plastic) the same color as the background, a sharply contrasted color from the background, the same color but different hue (say very light background and very dark rocks) and a background of one color with rocks of another but with similar shading to see if the contrast or color produces a different pattern.
 
Thanks for replying! The reason I want to keep them on each background for so long is to observe their behavior when surrounded by each different color. For example, does yellow make them hide all the time? Does red make them more active? I thought a week would be enough time that the octopus wasn't just having a strange day, but to actually see if each background affected it behaviorally.

Changing the rocks is an interesting idea. If I could get rocks that are the same color as the background that might be the best way to see how they match each color, since there would be nothing else to adapt to.

I'll look in to Dr. Hanlon's camouflage types, thank you.

Does anyone else have ideas or input? The more people that shade their thoughts the better!
 
I missed that you were looking for behavior and not physical patterning (however, patterning would still give you one quantifiable observation). Quantifying behavior is going to be difficult and observing it in an octopus doubly so. Researchers have been known to tear their hair out trying to get a handle on octo behavior. I suspect (now I am thinking a week may be overly short) you will need a motivator like food to introduce a behavior observation. Thinking along the lines of different food lid color with all the same jars. You could also use background matching and contrasting lids to see if they distinguish between colors. You would like have to get them all opening jars in their "regular" homes first then see if you can tell a difference in approach times and opening times. If you could train them to opaque containers, even better.
 
shipposhack,

Very interesting proposal. I hope that it goes well. If you want to quantify behavior in a statistically valid way, you will need to rigorously define the behaviors you're interested in beforehand. Obviously body patterning is an interesting behavior to note, but if you're interested in other aspects of behavior (e.g. activity or hiding) you'll want to be pretty clear about what constitutes these behaviors. For example, if an octopus is completely tucked away in its den, it would be pretty safe to define its behavior as hiding. But what if it decides to sit half way into its den but still partly sticking out. Is that hiding behavior? Or just sitting near its den?

To get an idea of the behaviors you are likely to observe (and how to define them) I would recommend examining an ethogram (catalogue of behaviors) for other octopuses. Christine Huffard published this one for Abdopus aculeatus (Ethogram of Abdopus aculeatus (d'Orbigny, 1834) (Cephalopoda: Octopodidae): Can behavioural characters inform octopodid taxomony and systematics?).

For the experiment itself, consider using the freeware JWatcher which is software designed to quantify observations of behavior. You could measure the frequency of behaviors (e.g. inking if they really don't like a color!) and the duration of behaviors (e.g. hiding).

Your technique for analyzing your data will be dependent upon your method of collecting data and the questions you are interested in. Are you interested in the frequency of behaviors? Duration of behaviors? How behaviors change over duration of time exposed to the color? The questions you are interested in will determine what statistical analyses you will perform. If you give a more precise idea of them, perhaps we could figure out the proper techniques to analyze the data.
 
Hi Nick,

I am a bit confused by what you are trying to do, and why. What is the hypothesis that you are trying to test here? And why are you planning to use different colors when it is fairly well recognized that cephalopods have only a single visual pigment (which means they are effectively totally colorblind?). If you are taking wild animals into the lab, I think you should have a strongly novel hypothesis behind you to justify that - I think the 'do cephalopods see color?' question is quite well answered already. Perhaps you could focus on the intensity/brightness of the backgrounds, using greyscale only? Not only does it give you a more controllable variable, but its also more supported by previous data.

If you do proceed with color, you will need to acknowledge that it's not actually the hue the animal sees, but the intensity. You will need a way to measure the monochromatic characteristics of the backgrounds you are going to use if they are not greyscale (you can create greyscale intensities easily, with color you will need to measure it to know what the animal will see in terms of brightness). You can probably also eliminate some of the colors you're planning to use because they are similar in intensity.

Regarding data and methods - I imagine one week is more than enough to get representative behavioral observations. I don't think they particularly like bare-bottomed tanks, but if you provide adequate shelter they can probably live with it.

You should also develop a firm protocol for how and when you will observe or photograph the animals. Octopuses are not overly fond of being photographed, so you need a way to ensure you are capturing representative behaviors when you take your photos - a camera that takes a photo every hour, for example, not a handheld camera snapping away when the animal does something 'interesting'. Likewise with your video observations - as Matt suggested, above, make sure you define your behavioral variables very tightly, and keep your observation periods to consistent times and situations. You could quantify frequency, as Matt suggested, and also look at changes over time that might suggest differences in the habituation rates to different backgrounds. With six animals you will need to have quite large effects to get statistically valid outcomes, so I would recommend working up a very firm protocol for your data collection so you can minimise the unexpected variations (which are the enemy of significant effects).
 
@robyn I thought about saying something regarding the lack of color vision but there are suggestions that the skin can "see" color. I've only seen a couple of references to the physical aspects but there are several animals that might fall into this category. I have seen nothing that would suggest that the color sensing feeds back to the brain so there may be interesting observations. Getting a clear handle on why he is testing and what he can quantitatively observe needs more thought and further research on how they MIGHT detect color should be part of his proposal/introduction.
 
The skin opsin that has been definitively identified is genetically identical to the one in the eye, which means its spectral range is also the same. The skin may indeed be able to sense light and intensity, but it is very unlikely the skin senses a broader range of color than the eye. And as DWhatley mentions, there is no evidence that light sensors in the skin are networked to the CNS - our own work recording from peripheral nerves in a range of cephalopods strongly suggests this is not the case, although it is possible it occurs in species other than the ones we tested (although those were the same species where the opsin has been described in the skin). I agree that a close discussion of the 'color vision' question is warranted, but I imagine a study using different colors that does not address their brightness and reflective characteristics is not going to be publishable.
Nick - can you tell us a little more about why you want to study color and what your hypothesis is? Maybe we can help with concrete advice on how to design your study to give you the most interesting data for your question.
 
Thanks for replying everyone! I have a couple professors here helping me with the project, but they don't know much about cephs, so it's good to get feedback from people who do.

Part of my idea for this project came from the last O. cyanea specimen we collected and cared for. After we caught him we transferred him to a yellow bait bucket while we drove back to school. He was obviously traumatized, but I noticed he remained pale and barely moved for the long car ride. We got him acclimated and put him in his tank, and he slowly became more friendly. One day I was wearing a yellow shirt very close to the same color as the bait bucket. The octopus was out playing and a few other people in the lab were there interacting with him. Every time I would walk near the tank (or within the octopus' vision) he would quickly run away and remain hidden until a minute or so after I left. I came and went many times, and every time he saw me he hid. I hypothesized that he might not respond well to yellow, either because the bucket we took him home in was that color and he remembered, or some other reason. This intrigued me.

I thought if there was a correlation between the bucket and shirt color and the octopus' behaviour it would be interesting to test. First I wanted to see how well the octopus could match the color, and next if it would act any different due to varying colors and patterns. It makes sense to use different hues of varying intensity. If you think this would provide better results I would be happy to change the colors to greyscale images.

I would like to test frequency of behaviour and how behaviour changes over time when exposed to the different backgrounds. My purpose was not to find out whether octopuses can see colors, but rather how well they match them despite being colorblind.

I look forward to more feedback!
 
I've been thinking more about my project and I had an idea. I was thinking about creating a map of different sections of skin that the octopus could control independently. I was watching a video from Roger Hanlon and he did this with cuttlefish. He found 13 (if I remember right) different sections on a cuttlefish that it can adapt individually. Has this been done before for octopuses? This would be a bit different than the project I was previously planning on, but I think it would be more beneficial and not as complicated. I could still do the same general plan, but with greyscale images and patterns.
 
I think this may be more viable and you have a good set of references to draw from. Alternately, if you can classify the same patterns that Hanlon's group feel may be universal, you might think about when the patterns are used (if any - this may be a no conclusion paper). You would have to set up quantifiable parameters for something like hunger, fear, relaxed that are not skin related and are repeatable but it would kind of meld the ideas you are refining.
 

Shop Amazon

Shop Amazon
Shop Amazon; support TONMO!
Shop Amazon
We are a participant in the Amazon Services LLC Associates Program, an affiliate program designed to provide a means for us to earn fees by linking to Amazon and affiliated sites.
Back
Top