Not so natural selection – The Selectinator 3000!

Evolutionary theory teaches us that when there is selection, a population will move towards a certain direction in response to this selection. Take for example the size of a dog. Generations of selection for small size finally led to a dog the size of a guinea pig, the Chihuahua.

When we people select for desired traits we call it artificial selection. Scientists do it all the time. However, when scientists do it, they are interested in creating different selection lines which they can use to answer a certain question. Selection for antibiotic-resistance in bacteria provides scientists with bacteria that are not killed by a certain antibiotic. Research in this line can teach us how bacteria are able to survive in the presence of this antibiotic. In my work I found that eggs of the flour beetle (Tribolium castaneum) without a serosa (more info here) have a lower survival in dry circumstances at lower temperatures (see figure below).

Insects develop slower at lower temperatures, so the simple explanation would be that the eggs are exposed longer at lower temperatures and so have more time to succumb to dehydration. Unfortunately there are more factors to take into account when looking at different temperatures, evaporation for example is faster at higher temperatures. To test whether the time in dry conditions is the determining factor for survival, we would need to select for fast and slow developing eggs so we can compare them at the same temperature. Selecting eggs for their development time does mean that I need to know how long each egg takes to develop and take the fastest and/or the slowest ones to the next generation. As the eggs vary about 30 hours in their developmental speed (see figure below), I would need to sit next to a batch of eggs for 30 hours and see when each egg hatches. This is of course impossible for larger numbers and multiple generations. Luckily my dad knows his way in electronics and together we build a selection machine that can separate larvae according to the time it took them to develop!

The selection machine – aka the Selectinator 3000

The principle of the selection machine is very simple. We use a piece of gauze with 300 µm holes in it. This is just small enough to prevent the eggs from falling through. A freshly hatched larva however is thin enough to squeeze through and will quickly crawl down. Once through the gauze he falls into a funnel down into a container below (see picture below).

The trick

The trick with this setup is that we build a machine that turns to the next container every 2 hours. So all the larvae are separated into different containers according to which 2-hour slot they hatched! This particular machine has 3 different gauze/funnel/container combinations which means we can measure the hatching time of 3 different batches of eggs simultaneously. See the movie below to see the machine in action, I made it turn a couple of times for the movie, normally it turns every two hours.

The first results

I normally don’t post unpublished data on my website, but this time I make an exception. Below is a figure with the time it takes eggs to develop for 3 different lines after 9 generations of selection. 1 line is selected for fast development (in red), 1 is not selected (in green) and 1 is selected for slow development (in blue).

As you can see, there are clear differences between lines! In fact, the peak in the number of larvae hatching is at 133 hours in the fast line and at 153 hours in the slow line. That is already a 20 hour difference! This also means that I can compare the hatching rates at low humidity and see whether the slow developing larva have a higher mortality than the fast developing ones. And this without the other factors that are involved when using different temperatures.

What’s next?

So these are some preliminary results. For now I will keep on selecting and try to increase the difference between lines. Once I have a satisfactory difference I will test my hypothesis. This will surely provide us with some interesting insights into the evolutionary process and will likely be useful for other scientists too that have questions related to developmental speed.