Thread by @BiolBulletin

It's #ReadAlong Day!

@AnnTarrant2 will #TakeOver our account at 12 pm ET.

Follow this thread for live tweets as we #ReadAlong: Clarke, 1933 (journals.uchicago.edu/doi/pdf/10.230…)

Drop your questions and comments in this thread.

Happy Reading! 🧵(1/n)

https://twitter.com/BiolBulletin/status/1467213652990349315

.@AnnTarrant2: Hello! Thanks for joining me on this #ReadAlong with @BiolBulletin. My name is Ann Tarrant, and my lab (www2.whoi.edu/site/tarrantla…) studies how marine invertebrates detect and respond to environmental signals and stressors. (2/n)

I'm a Senior Scientist at Woods Hole Oceanographic Institution (@WHOI) and the (newly-minted) Associate Dean of Academic Programs. Ask me if you have questions about our educational programs! (3/n)

About half the research in my lab focuses on #copepod physiology. One of my recent projects (more about that later) is focused around diel vertical migration, the subject of today's paper. (4/n)

Diel vertical migration (DVM) describes the movement of animals through the water column, usually into deep water during the day and shallow water at night. Here's a fun animation of DVM (© @WHOI). (5/n)

I was asked about my approach to reading papers. It honestly depends on my purpose for reading. Am I trying to learn about a new area? Get help on some methods? Dig into results within my area of expertise? (6/n)

Usually I start with the abstract and introduction...I want to understand the author's "world view" and for modern papers, I want to gain context on the state of the field. (7/n)

Then I usually "hopscotch" through the paper, focusing on understanding what's in the figures and digging more deeply into the text depending on my level of interest. (8/n)

After that, I usually read the discussion more thoroughly. I try to soak up the big messages and how it relates to the larger body of work....but because I often skim the methods and results, I end up flipping back and forth to really understand the discussion. (9/n)

For today's read-along I've decided to dig waaay back to 1933 and discuss some old-school work by on DVM by another @WHOI scientist, George L. Clarke (p.s. he worked at @Harvard too). (10/n)

The introduction is pretty short and sweet. Clarke references previous studies of DVM (which he calls "diurnal migration"). While light seemed to be the most important cue, scientists didn't understand precisely how light levels related to migratory patterns. (11/n)

As an aside: you gotta love the old-school spelling of zoöplankton! 😉 (12/n)

Today we know that small animals migrate to avoid being seen (and eaten!) by visual predators. They hide in deep water during the day and move up at night to eat phytoplankton, and their predators follow. This was all apparently a bit mysterious in back in 1933. (13/n)

At the time, researchers obviously didn't have the same instrumentation that we do today. Clark proposes to make two improvements on previous studies. (14/n)

First, he measures irradiation while sampling - continuous measurements on deck and vertical profiles "at suitable periods." (15/n)

Second, he worked with "Mr. C. O'D. Iselin" (Columbus Iselin, @WHOI's second director) to build a contraption to simultaneously sample zooplankton from 5 depths (more on that later). Today, that contribution would probably rate co-authorship for Iselin. (16/n)

Columbus Iselin: en.wikipedia.org/wiki/Columbus_… (17/n)

The intro finishes by explaining that: Observations were made from the "Atlantis," which was brand-new at the time. The original Atlantis (1931-1964) was followed by the Atlantis II (1963-1996), and a "new" Atlantis (2007-present). (18/n)

A lot of things have changed since then. The Atlantis used to have cats on board! Here's "Nosey" investigating the lines. I definitely vote to bring back the cats. (19/n)

Turning back to the methods used in this paper, they had to build the net system used to sample. They even had to make the next themselves...silk didn't work too well, so they used coarse cotton scrim (a gauze-like fabric) (20/n)

They basically laid out 5 nets along a 50 m-long cable. The nets were kept closed on the way down, opened for the tow, and then were mechanically closed before pulling the whole string up. (21/n)

Seems amazing that this thing worked and didn't end up as a jumbled-up mess! (22/n)

If that's not crazy enough, to get a 100-m profile, they deployed the thing twice (0-50 m and 50-100 m). (23/n)

Here's a photo (circa 1932) of Clark deploying the net system. The nets were attached to the cable from this platform. (24/n)

The mechanics of the whole thing are clever and frankly a little over my head (I already admitted that I tend to skim the methods). It sounds like there are lots of "fiddly bits" that had to fall into place... (25/n)

If you are interested in the evolution of sampling equipment, I highly recommend Wiebe and Benfield 2003: sciencedirect.com/science/articl… (26/n)

....by the way, I'm trying to make this fun even if you haven't read the paper (I won't tell if you don't!) 🤫(27/n)

Moving on now past the net design to the actual sampling (second paragraph of p. 407). They measured the cable made with vertical "in the usual way" (I love this). They needed to know this in order to calculate the depth of the nets. (28/n)

Here's 1950s picture of Fritz Fuglister measuring a wire angle...we use similar instruments today! (29/n)

And yes, oceanographers do need trigonometry! (30/n)

The next paragraph kind of sounds like a word problem to me. To calculate the density of animals, they needed to know the ship speed and then could calculate the volume of water filtered... (31/n)

They did this by throwing a piece of wood into the water and timing how fast the ship moved past it. Cool! (32/n)

Top of p. 408 describes the tows they completed. It's AMAZING they got the system to work so reliably. (33/n)

Now we get to "Sources of Error" (p. 408). One issue is patchiness of the plankton due to "swarming." Clarke largely discounted this (perhaps too much) based on the distances over which they towed and high repeatability.(34/n)

I like the "Sources of Error" section. It kind of reminds me of my high school lab notebooks. But it's also a little self-congratulatory...they seem to think the errors are pretty minor. (35/n)

Other issues mentioned: limited depths over which they could sample, clogging and escape. In particular a potential bias is that it might be easier to escape when light levels are higher. (36/n)

Today, strobe lights 😵‍💫😳🥺💡 are sometimes used to prevent animals from seeing and escaping from nets. Plankton pumps also avoid some of the biases associated with net sampling. (37/n)

Now in "Observations" (p. 411). Could you imagine making Fig. 3 by hand?! I'm guessing Clarke would have been a big fan of geom_violin()! The plot "constructed in the usual way" compares daytime distributions across species and doesn't yet address patterns of DVM. (38/n)

One pattern is that adults and older stages tend to be deeper than the smaller younger stages. This might reflect differences in vulnerability to predation. (39/n)

Metridia lucens, one of the focal species produces bioluminescence! (Metridia sp. Shown here) (40/n)

It's a little weird in Figure 3 that Metridia females are on a different scale (because there are sooo many), but they don't show the scale. (41/n)

You'll also notice in Fig 3 that most of the Calanus large juveniles ("Cop V") are down deep. That observation is actually a little tricky to interpret... (42/n)

It could be diel migration, but these copepods also undergo a seasonal dormancy in deep water. In the Gulf of Maine (study region for this paper) juvenile copepods enter dormancy in summer/fall. Many of the juveniles in deep water were probably dormant. (43/n)

Then come 8 PAGES of tables! ⚠️This was obviously before the days of online supplements. It's nice to have all the information in one place, but the supplements allow us to provide more information, often in usable formats. What do you think? (44/n)

On p 420, Clarke continues talking about Fig 3. Centropages tend to be shallow. These are the smallest of the 3 species shown and the least vulnerable to visual predators. (45/n)

There's a footnote (p. 420) acknowledging Miss Mildred Campbell, who quantified stages of juvenile copepods in a couple of tows. (46/n)

BTW: Mildred Campbell was a bada** copepodologist in her own right! She spent time as a @WHOI research assistant, and earned her Ph.D. from the University of Toronto (@UofT) shortly after this paper was published. More on Mildred: doi.org/10.1651/11-350… (47/n)

Next (p. 421, tables IV-VI) they measured copepods from each tow/net. It took a moment to figure out that Tables IV-VI are frequency tables. If I were reviewing this, I would have asked for more informative titles! (48/n)

Clarke doesn't really explain that the size of a copepod within a growth stage depends on its feeding history and especially on the water temperature during development. Colder water leads to slower growth and larger body size. (49/n)

Clarke argues that within a stage and species, the larger copepods tend to be in deeper waters. This is most apparent for Calanus (Table VI... with a very unfortunate page break!). (50/n)

For Metridia, it wasn't all that obvious to me. There seems to be some trend when you compare shallow nets (32-42m) ~1.96 mm avg with deeper nets (54-114 m) ~2.05 mm, but I'm not totally convinced. (51/n)

On p. 423 "If the difference be regarded as significant..." Is it? 🤷‍♀️ I don't know... you tell us. A little more quantitative analysis would have been helpful. (52/n)

The physical-chemical measurements in Figure 4 were made by "Dr. Redfield" ...yes, that guy, the creator of the Redfield ratio (106:16:1, ratio of C:N:P in marine phytoplankton) [en.wikipedia.org/wiki/Redfield_…] (53/n)

My lab is actually in the Redfield Building 😁 (54/n)

Fig 4 shows a thermocline around 20 m. They couldn't measure phytoplankton (food!) until ~3 weeks later... today we would be able to make simultaneous measurements (at least with a fluorometer on a CTD) (55/n)

The main phytoplankton they observed (bottom of p. 424) were types of dinoflagellates and coccolithophores. Of course they weren't able to observe very small (and incredibly abundant) cyanobacteria at the time. (56/n)

With Figure 5, we finally get to the point of the paper! Here, Clarke maps depth distribution of female Metridia in relation to subsurface light levels over a 40-hour period. (57/n)

He was particularly interested in showing they seem to follow specific light levels (the curved lines). (58/n)

In general Metridia density was shallowest just before dawn, when they came up to right around the thermocline... though the scientists seem to have taken a few hours off from sampling in the wee hours of the night.😴😴 Believe me, I understand! (59/n)

They made an interesting/odd observation of sharp drop between ~4 am and 4:41 on the first day. They don't know why.

...maybe it's that patchiness that they said wasn't a problem back in "Sources of error" 🤦‍♀️
(60/n)

They suggest (bottom p. 427) that in some cases copepods might not be able to swim fast enough to match light changes at dawn/dusk. (61/n)

They calculate that the copepods would need to swim 77 cm/min to follow the most rapid changes in light. One study cited (Welsh, 1933 @BiolBulletin; journals.uchicago.edu/doi/pdf/10.230…) suggested that it should be possible. (62/n)

(Bottom of p. 428) migration by Centropages was less pronounced. Clarke doesn't really speculate on why... today we would propose that because they are smaller, predators can't see them as well, so less pressure to migrate. (63/n)

Might as well stay in shallow water where the food is 🥦🥕🍕(64/n)

Calanus distribution also didn't show much day/night movement... as I mentioned, I think this is because many were dormant. (65/n)

Also, Calanus has large lipid stores in their body. If conditions aren't good for migrating to the surface, they can afford to "wait it out" for a couple days. (66/n)

Clarke continues to discuss detailed observations of day/night patterns, but we hit the main points... moving on to discussion... (67/n)

This study agrees with previous work showing differences among species, developmental stages, and "broods." (68/n)

I think "broods" are cohorts, meaning behavior changes over the year. That could relate to predator abundance, food availability and dormancy. (69/n)

Clarke and others he cites (e.g., Bigelow 1924) seem baffled by variability in Calanus behavior. I've been studying Calanus dormancy since ~2003, and I'm still kind of baffled. (70/n)

...we know a lot more about the timing of the phenomenon and ecological drivers, but still can't reliable induce dormancy in the lab! (80/n)

I love the Bigelow quote that they sometimes "made rich catches [of Calanus] on the surface when the sun was high in the sky." (81/n)

And (bottom of p 431) Clarke grandly states "Let us therefore continue our search for a general explanation to those organisms which do migrate.."

That's some gravitas for you! (82/n)

The major point is that Metridia female distribution seems mainly linked with light levels, but there seem to be other factors at play... (83/n)

Clarke suggests food availability, the position of the thermocline, and gravity (geotropism) may be important. (84/n)

Today, we might also consider circadian rhythms as drivers of this behavior and daily changes in the copepod physiology (how hungry they are!) (85/n)

Clarke also points out (p. 433) that it's not quite clear whether the copepods are responding to absolute light intensity or *changes* in light levels. (86/n)

He argues that we need more lab experiments to test theories about DVM. On the one hand, I love a good lab experiment... (87/n)

But on the other hand, maybe he's giving up on field studies too easily. (Though maybe that's easy for me to say... field work is HARD and was even harder back in the day) (88/n)

The paper ends with a pretty nice summary. I agree with points #1-6, but the last point seems a little too strong ("This investigation confirms the idea that light is the most important factor controlling diurnal migration"). (89/n)

Light is the most important physical factor, but they have completely neglected predation as a driver. (90/n)

Do you ever wonder what people will say about your papers ~90 years from now? Kind of scary, huh? (91/n)

That wraps up my thoughts on Clarke 1933! Thanks so much for joining me 🥰 I'll reply to any comments as @AnnTarrant2. (fin)

Please unroll @threadreaderapp