Sellnickia: Details & Enhancements

Sellnickia anterior with details: sp. A, sp. B

Here’s some more on the predatory labidostome mite from the previous post – a closer and greener view of the anterior end. It’s greener, because I distinctly remember a live collection of a species of Sellnickia with unusually greenish mites running around, many with startlingly white Folsomia-like springtails crushed in their chelicerae. A reliable memory or a dream? I do dream of outlandish mites every now and then, some of my favourite dreams I might add, although my absolute favourite was when I discovered a giant trigonotarbid in a deep, misty, tree fern and liverwort covered canyon. Since trigonotarbids (allegedly) haven’t been around for a couple of hundred million years, I suppose this indicates one should be skeptical of their dreams.

In any case, this reminds me that Adrian has asked for more details on the time it takes to make these mite portraits.  Warning – what follows is lengthy.

To give even a general answer to ‘how long’, one needs to first decide how the image is going to be put to use. If the purpose is to illustrate morphology for a scientific publication, then the less time spent manipulating the image, the better. For example, the detail of the cuticle from sp. B in the image above was simply selected, copied and pasted. To illustrate why: I was once told a story about an early photographic plate of fossil aquatic scorpions. Although aquatic chelicerates typically have compound eyes (think horseshoe crab), modern terrestrial scorpions are typical ‘arachnids’ with at most lateral clusters of simple ocelli. No eyes were obvious on the aquatic rock scorpions in the photo plate, but I am told that the actual impressions in the rock have lateral compound eyes. Apparently, since scorpions weren’t supposed to have compound eyes, the ‘artefacts’ had been airbrushed out by the author so as not to confuse the reader. Once you start ‘improving’ an image, you run the risk of producing misinformation.

The time required to prepare a SEM for a scientific publication is primarily a function of specimen preparation time. For example, to produce a grayscale SEM suitable for publication of the Sellnickia sp. A above took less than 3 hours including selecting the mite, drying it through a series of solvents to eliminate its water content, placing it on a stub, sputter coating in gold, putting it into the electron microscope and achieving the proper vacuum, focusing and fiddling with astigmatism etc., and taking the images. Three hours is a lot of time to take one picture, so I always do lots of specimens on the same stub. Mites are excellent in this way – lots more fit on a stub (~10 mm diameter) than could dance on the head of a pin. So, 2-4 hours of preparation can result in 30-60 images in a 4 hour SEM run, depending on time needed to focus, resolution (the more resolution the longer the raster time for each image grab), software or hardware problems (all too frequent), and the quality of the specimens (finding specimens on the stub without dust, goo, dents, or broken bits).

The reasons this mite was easy are: (a) it is hard shelled (little or no deformation on drying) and relatively large (much easier to transport onto a stub) and (b) relatively flat (not much depth of field in a dorsal view).

The image in this post is from a single grab – if you look closely, some of the ornamentation and many of the setae are not in sharp focus and the blow-up of the detail (A) is a bit blurry.

The image of the full mite in the previous post is a composite of 4 – I divided the mite into quadrants and took four separate SEMs and then pasted them together in Photoshop – mostly to get a large image size (the camera was only capable of relatively low resolution grabs).

Masking – separating the image from the background – takes the most time. In the case of the full body image in the previous post, it took 7-8 hours to put together and mask – that is not bad and is a function of the outline of the animal – few setae and other protuberances to mask around.

In contrast, the anterior view in the image in this post took about an hour to mask this morning (including erasing the legs that were bunched up and out of focus on either side). Masking out artefacts (e.g. cracks and bubbles in the glue) or unwanted detail (e.g. the out of focus legs in this picture) can dramatically improve the quality of a greyscale image in a paper – but it takes a lot of time. It would probably be a good idea to mention in the legend of the figure that the legs have been removed.

Once the animal is masked, then you can decide it you want to colour it. I don’t think that a morphologial study needs colour, but a field guide or poster could benefit from colouring.  The problem with a grayscale SEM is that we have colour vision and when we see these mites they look bright yellow to green. Colouring can be simple – the full dorsal habitus view in the previous post took only about 20 minutes to do after masking, because I relied on the shade differences in the original SEM to be reflected in the final picture and used a single colourizing level in Photoshop. If I wanted to show more detail, then I would have to laboriously select the areas needing the different colours one by one.

I actually did this on this morning’s image – the anterior view in this post. I decided to give the mite a more greenish cast in parts of the cuticular design by selecting pixels within the reticula. I also decided to lighten the setae – the setae usually lack the colour of the body in life and appear white. This took way too long – especially trying to select the feathery bothridial sensilla that are overlain on the body. Finally, I decided make the tips of the chelicerae a slightly different shade. All of these changes should be making the mite look more realistic, or at least make it easier to see the different parts, but even these fairly simple selections took about an hour and a half.

So, the total time to prepare the image in this post was about two and a half hours on a Sunday morning. Getting the original image was a similar investment in time. I would consider this a low value for the average SEM that I colourize – the average is about one full day of manipulation of the original grayscale image(s). The maximum is about a full week – the extremely complex  image posted near thebeginning of this blog.