
			Society for Amateur Scientists Plants at Less than 1g

Growing Plants at Less Than 1 g

by Shawn Carlson

This monograph is supplemental information to "The Amateur Scientist," Scientific American, February 1996. You won't be able to make much sense of it unless you've read that article first.

How Well Does A Coventional Clinostat Mimic Weightlessness?

A clinostat does not mimic weightlessness exactly. The gravity sensors inside the plant's cells are confused by the constant rotation, but not blinded. They do sense that something is happening and they do promote an increase in auxin production. However, the continual rotation prevents the auxin from distributing along the bottom of the stems and roots as a simple constant tilt would do. Although growth studies done inside the space shuttle have not yet detected a difference between the way seedlings grow in space and in a clinostat, many biologists are skeptical that plants grown in clinostats will behave exactly as those grown in a long term in weightlessness, say inside a space station. To date, the research suggests that any differences between the way plants grow in space and in a clinostat are small, but just how small is still an open question. The answer can not be found on the ground.

Active Control Of The Wheel's Rotation Rate

As the wheel tumbles around, its weight shifts and the shaft sloshes a bit in its support. This causes changing frictional loads that can affect the rotation rate, and thereby the effective gravity. The difference in loading is a potential problem even if you were to mount the shaft on ball bearings. A professional-quality system requires some sort of active control that keeps the rotation rate constant.

The simplest way to do this is probably to replace the small high-speed motor with a stepping motor. A stepping motor rotates through a precise angle every time the motor assembly receives a voltage pulse. Suppose for instance, a particular stepping motor turned its shaft precisely 120 degrees every time it received a brief electronic pulse. Three pulses would turn the shaft 360 degrees or one complete revolution. So, for example, if you were to feed this motor pulses at a rate of 90 per second, the motor shaft would rotate 30 per second. Steppers are designed to move at whatever rate they are driven at no matter what their load (within reason) and so they are ideal for situations like ours where a variable load has to be driven at a constant rate. Steppers can't run too fast, about 500 rpm is the limit. However, by placing a wide-diameter drive shaft on the stepper, or using a little gearing, they can drive the bicycle wheel at the required speeds.

Stepping motors are available in kits from a variety of sources for under $20. Any well-stocked electronics store will have them as well as many mail-order electronics catalogues. You want to make sure to purchase the motor along with its controller card. The controller card is the electronics that actually drives the motor. It does all the hard part. Then all you have to do is supply the controller with logic pulses (square pulses of well-defined voltage and duration) at the right rate to drive the motor at the correct rpm.

You can supply the pulses by using two simple circuits each designed around the 555 timer. The 555 timer is the most commonly used timing circuit by hobbyists. It's a very versatile little integrated circuit chip and can be purchased for about a dollar from any Radio Shack and just about any other electronics supply house. The first circuit you will need is a simple clock that produces pulses at the correct rate. These pulses are fed into a second 555 timer that has been wired to function as a "bounceless switch"-- a circuit which produces a standard logic pulse each time it is triggered. The combination generates standard logic pulses at a desired rate which are then fed to the stepping motor. If you build a timing circuit that allows you to vary the pulse rate, you will be able to change the rotation rate of the wheel and so adjust the acceleration that the seedlings feel in your trials.

Any electronic book that describes the 555 timer will provide schematics for both a simple timer and a bounceless switch. All you have to do is figure out the rate at which you need to drive the stepper motor at, choose the correct resistor/capacitor combinations to drive the timer circuit at the desired rate, pass the output pulses into the bounceless switch and then feed it's output into the controller card. If you have a Radio Shack near by, get Forrest Mims' "IC Timer Circuits". It costs about $2 and tells you everything you need to know.

There are, of course, many other ways to control the rotation rate, but none are as simple or make it as easy to modify the rotation rate (and thereby change the conditions of the experiment) as a stepping motor.

Suggested Experiments

Here is a list of novel experiments you can do. Your work could provide original discoveries that are fundamental in understanding how plants sense and respond to gravity.

Measuring The Gravity Threshold

This is the fundamental experiment. Germinate seedlings at a number of different accelerations and identify the minimum acceleration needed for the seedlings to orient themselves to it. My original work used a quantity I call the "angularity"-- the sum of all the bend angles in the plant's stem. It's measured by cutting each seedling at each bend and then laying the pieces together with all the bends going in the same direction. The angularity is then simply the angle between the first and last piece.

If you measure the angularity of seedlings germinated near 1 g, the angularity is quite small with small variance. (The variance, also called the standard deviation, is a measure of how spread out a group of measurements is. A detailed description of variance is outside the scope of this monograph. If you don't know about variances, get a good book on statistics for experimental scientists and learn all about them! These statistical tools are absolutely essential for any researcher-- professional or amateur.) If you measure the angularity of seedlings germinated near zero g, their average angularity is usually much larger and so is the variance. In "The Amateur Scientist" article, I suggested using the average value of the angularity to try and identify when the gravity sensors inside the seedlings kick in. This works but there is often a problem. You need to be able to draw a curve that traces the angularity verses the acceleration. It is the shape of this curve that tells you where the gravity sensors become active. The problem is that the variance is sometimes so large near low accelerations that it is often quite difficult to know where the curve should be drawn. You can sometimes draw any number of curves and the data doesn't distinguish very well between them.

When this happens I have had better luck by ignoring the average angularity and instead graphing just the variances vs. acceleration. If you crack open a statistics book you'll generally be given advice like this: Measure the variance for say 30 seedlings at a given acceleration and use that as the quantity of interest. Then repeat that measurement 10 times at the same acceleration so you can get a feel how the variance varies at each acceleration. That is to say, you want to take the variance of the variance of the angularity. (Don't get confused now. It sounds strange but it's really straight forward if you go through it slowly.) Then graph the variance of the angularity vs. acceleration. The variance of the variance is the number you use to draw your error bars.

Keep in mind that if you can see how gravity turns on inside the plant using just the angularity measurement, then USE THE ANGULARITY! The variance method requires you to take more data and do considerably more work. However, for each plant you should try both methods at least once so you can compare your results. If the two methods give statistically inconsistent results you will want to report both. And always be sure to let your colleagues know which method you used.

In what follows to make this presentation clear, I'll assume you are using the variance method. You can use the same ideas to analyze the angularity curves as well.

You should find that the plot of variance vs. acceleration is easy to interpret. It's high at low accelerations and low at high accelerations. The inflection point, the point in the middle where the curve starts to change from a downward bend towards an upward bend (for you Calculus buffs, it's the place where the second derivative of the variance function goes to zero) makes a convenient spot to define experimentally. I refer to this simply as 'a' subscript 'VIP' which stands for "variance inflection point." The lower this number is, the smaller the acceleration threshold that the seedlings can detect.

You can pick out this point pretty accurately by eye. You purists will want to perform a parametric fit to the data. (We use a parametric fit since no model is available to explain why the data should look the way they do). Then take the second derivative of your fitting function and set that to zero and solve for the acceleration.

It is important to note that nothing interesting may be going on inside the plant at the acceleration corresponding to the variance inflection point. Clearly, the sensing mechanism must kick in at an acceleration lower than the VIP. Again, since we don't have a mathematical model that explains our variance curve we choose the VIP merely as a convenient point for our fellow experimenters to reference.

The variance curve itself has two important features for the experimentalist; the inflection point, and the range in acceleration over which the plant's response to gravity changes-- that is to say the width of the transition region between no response (low gravity, high variance) and strong response, (high gravity, low variance). The width of this region can be quantified in all sorts of ways, and in the absence of a mathematical model that lets us calculate the shape of this curve, one way is as good as any other.

Here's how I do it. Let V(0) be the variance measured near zero acceleration (that is, very near the axis of the bicycle wheel) and V(1) be the variance measured at 1 g. Then the transition region takes us from V(0) to V(1). I define the lower bound of the transition region to be the acceleration at which the variance equals V(0) - (V(0) - V(1)x e ^-1. I like the e^-1 term (that's the number e raised to the -1 power. e is an irrational number that pops up all the time in physics. e^-1 = 0.37 approximately) only because I'm a physicist and in physics e appears almost as often as pi. I define the upper bound of the transition to be the acceleration at which the variance equals V(1) + (V(0) - V(1)) x e^-1.

Of course, to make this measurement takes a bit of work. At the very minimum you need to grow 30 seeds at each of 5 accelerations to measure the curve and three of these points have to be on the transition part of the curve. If you limited yourself to just five accelerations you would have to be very lucky to hit the right accelerations. Realistically, you'll need to do 10 accelerations to have a good change of defining the behavior. That means that you have to grow, cut up and measure about 300 seedlings to get a reasonable sense of the variance curve. If that seems like a lot of work, well, welcome to the world of science. Nature rarely gives up her secrets easily. I remain inspired by reminding myself that no one in history ever knew what I am about to know, and that my efforts will soon contribute to our understanding of the world. For me, the thrill of discovery makes all the careful work worth while.

There are a couple of things you can do to make your life easier. First, build several bicycle clinostats and set them up so that only a manageable number of seeds become ready for processing each day. You can easily run each with 30 seedlings at each of five accelerations (at five different distances away from the center of the wheel). Harvest one clinostat each day until you have completed your run.

A second thing you can do is find other people who are willing to share the labor in order to share in the fun of discovery and in the joy of making a contribution to knowledge. Collaborations are a fundamental part of science because the amount of work needed to make a discovery is often simply too much for one person to handle.

If you've got too many seedlings to process immediately place the extra in the crisper box in your refrigerator. That stunts them in the sort term. You will still want to process them as quickly as possible. I've thought about freezing the seedlings so I could process them over several days, but I'm concerned that the freezing will disrupt their cellular structure and cause the seedlings to fall limp and thereby alter the angularity measurement. I invite someone out there to do an experiment and see if freezing the seedlings significantly alters the angularity. Please let me know what you discover so SAS can share it with other researchers.

Alert readers may have caught a mathematical error when I said that it takes about 300 individual seedlings to map out the variance curve. Earlier I said that it takes about 30 seedlings to make a good measure of the variance and that it takes about 10 measurements of the variance to get understanding of how the variance changes (that is again, the variance of the variance). That would come to 300 seedlings per acceleration or 3000 seedlings per variance curve! Science is often a balance between doing what the statistics books tell you you ought to do and what is practical. Since these variance curves have never been measured for most plants, you can make discoveries even if you do not take the amount of data normally advised in statistics books. To make progress rapidly, you want to take the least amount of data you can and still get a reasonable sense of the quantity your measuring. If you are trying to repeat someone else's measurement you'll need to take at least as much data as your predecessor did. However, if you are in uncharted waters, just about any experiment you do carefully will result in a discovery. My experience is that you can do a fairly good job at determining the variance curve with 300 seedlings. Use that number until you discover experimentally that you can get away with fewer or that you need more.

By growing seedlings and plotting the variance as described you can make important discoveries about how plants evolved to use gravity to their advantage. Make sure you share your discoveries with the Society for Amateur Scientists so we can help you get them published in a professional research journal so you can share them with the rest of the world!

Other Experiments

Does The Variance Curve Change With Germination Time?

Germinate the seeds of a single species for different lengths of time and construct the variance curve for each. Does the variance curve depend on the time the seed was allowed to germinate? What does this say about the development of the gravity sensing mechanism? If a development effect exists, does it correlate with physical features of the adult plant? For instance, do plants which grow large tend to develop their gravity sensing mechanism at a different rate than plants that remain fairly small?

Behavior Of Plants Within A Genus Or Family

Take a number of related species of plants within a single genus and compare their variance curves. How similar are they? Next try the same thing with plants that are not is the same genus but which are in the same family. Has the gravitation sensor evolved over time?

Can Gravity Threshold Be Related To Other Traits

Can you identify any physical traits of those plants with low VIPs share? What about plants with high VIPs? Actually, I would be surprised if any such traits show up, but hey, what do I know it. It's prejudices like mine that may prevent me from making an important discovery, simply because I spend my time looking in areas where I suspect I'm more likely to find something interesting. If you look for such correlations you might make a discovery.

Distribution Of Angularity

Histogram the value of angularities you measure for each set of conditions. Is the histogram "bell shaped" for all accelerations? If not, why not? Are you looking at something important in the plants structure, or is what you're observing merely statistics in action? Make sure to look at the angularity measured near 1 g.

Planting Seedlings And Charting Subsequent Growth

Once you've measured a species' VIP vs. the number of days it germinates you can experiment with how well it recovers vs. time spent in the clinostat. Grow a set of control plants-- plants you do not expose to low gravity-- for each trial. Begin by exposing 30 seedlings to low gravity for 1 day. Then place these in the same environment in which you are growing your control seeds. Note any difference between the developments of the two groups. Do the same for plants exposed to low gravity for 5 days and 10 days. Measure growth rate, leaf development, sexual development and other physical features.

Can you identify any difference in cell structure? If you crush some of the plants up can you find any evidence of chemical substances that are in one group but not another? (Describing how to look for and how to carry this out is beyond the scope of this monograph. If you are interested in pursuing the biochemistry of your plants consult some laboratory books on botany and biochemistry for ideas about what to do here.)

Auxin Levels

Just how much auxin is produced over time inside a clinostat and how much remains in the plant's tissues is an open question for nearly all plants. Here also is a great avenue for amateur exploration. Unfortunately, the methods of measuring auxin levels are beyond the scope of this monograph, but there are many good books on plant physiology that you can look up to learn how to make these measurements.

How Does Auxin Production Depend On Tumbling Rate?

Measure the auxin levels in seedlings over time as they are exposed to sub-gravity environments. Pick at least 5 different accelerations to expose your plants to and see how much auxin collects in their roots and in their stems over time. Then vary the rate at which you tumble the bicycle wheel and do it all over again. Do your plants produce more or less auxin as the rate is increased or decreased? Plot the auxin levels vs. tumbling rate.

This experiment is important in understanding how well these sub-gravity clinostats simulate weightlessness for the plant. If the auxin levels remain high throughout the plant's time in the clinostat, skeptics can fairly question the applicability of this technique to simulating weightlessness.

After you've done one plant, try other species. If you can find one which does not produce abnormal auxin levels during your trials, then that plant could be a true window into geotropism.

Additional Experiments

There are all sorts of interesting things to look at. If you've got ideas for projects not listed here, please tell us about them. We'll post them to the SAS Web page and give you credit for the suggestion.

Slip Rings

Our clinostat uses a dual slip ring assembly to bring power to the motor that drives the wheel. These parts can be difficult to come by. There are lots of companies that sell suitable dual slip ring assemblies; to find one near you, go to your local library and look up "slip rings" in the Thomas Registry of manufacturers. However, these devices can be fairly expensive when purchased new. On the other hand, home-made slip rings tend to produce a lot of electrical noise due to fluctuations in contact resistance as the assembly rotates. As a result, they are rarely used when small voltages are being transferred to or from a rotating platform. However, in our application, this slip ring assembly is only being used to provide power for the motor and so with a little electronic buffering after the current is pulled through the ring, (the circuit diagram is provided in The Amateur Scientist article) we can use an inexpensive home-brew slip ring assembly.

Building a good slip ring assembly is labor intensive. But for you purists with a lot of time on your hands, I'll describe what may be the easiest way to build our dual-ring assembly from scratch that works well. This is, in fact, the arrangement that I use. However, let me first caution you against wasting your time trying to adapt a ball-bearing assembly to do the job. Even with a conducting lubricant, like carbon powder or salt water, I have never been able to produce a unit with resistance fluctuations less than 100K during a single revolution that performed well for longer than a few hours. (And boy how I have tried! Mercury may work better, but the stuff is too dangerous to handle so I've never tested it.) If you know how to adapt a ball bearing assembly to be a low-noise slip ring, please let me know and I'll publish your solution on the SAS Web page.

Simple Slip Ring Assembly

You will need to install these slip rings before you install the shaft that carries the counter weight.

Sand off the original paint and any rust that may have accumulated on the shaft on the head tube between the two vertical wood supports (see diagram in The Amateur Scientist). Then insulate this length by painting the entire shaft with latex-based enamel paint. Drill two small holes, each about an inch and a half from either support, large enough to thread an insulated wire through. Next, paint the insides of these holes to insulate the exposed metal. If you don't take care to keep all these surfaces well insulated this slip ring assembly will develop a short circuit.

Drill a somewhat larger hole through the shaft under the electronics box. Thread a length of insulated wire from each of the small holes to the large hole. It's easiest to insert wire in the small hole and to sang it near the larger hole. A hook bent out of a paper clip makes an excellent tool for this job. Make sure you use different color wires so you can easily tell which is which. Trim the wires to leave at ten inches of wire dangling out of the small holes. Strip the ends protruding through the large hole and connect the wire going to the near hole to the positive input of the power circuit and the wire going to the far hole to the ground input.

Strip each of the wires so that about one half inch of insulation extends above the small holes. Next, cut two strips of aluminum foil each about two inches wide and ten inches long. Puncture one strip at the center line about one inch from the end with the bare end of the wire nearest the bicycle wheel. Thread the foil down the wire until it just rests on the head tube's shaft. Then bend up the end of the foil and paint it with latex based enamel paint. When the paint dries, dope the painted area with epoxy and gently press the foil over the hole making absolutely certain not to push any aluminum into the hole. The insulated part of the wire should poke through the aluminum foil.

After the epoxy has set, tightly wrap the foil strip and the bare wire together around and around the shaft. As you do so, dab a little epoxy at least a quarter of an inch away from the wire on either side of the wire to hold the layers together. You don't want the epoxy to touch the wire. Seal the end of the strip with a liberal dab of epoxy.

Next, pinch up the sides of the strip to create a gentle 'U' shape across the strip. You want the sides to be slightly higher than the center.

Repeat this entire procedure for the second aluminum strip and the second wire. At this point the assembly that takes the interfaces with the non-rotating world and passes the current to the motor assembly is complete.

Check for short circuits with the continuity tester on your digital volt meter. If the continuity test is positive for a connection between the two aluminum strips, now is the time to take it apart and rebuild.

To bring the power to the this assemble, install two short wood screws on the slat which joins the two vertical supports (see diagram in The Amateur Scientist) so that each one is in line with the center of the of the aluminum straps you have just stalled. The screws should be at least one inch above the strips and facing in the direction that the motor mounted on the outer vertical support rotates the wheel. You are about to install a wire from the screw to the shaft and you want the shaft to rotate with the wire, not against it.

Run two wires from the power strip that goes to the AC to DC adapter up the inner vertical support. Connect the wire that goes to the positive DC terminal to the screw opposite the inner strip and the other wire to the screws opposite the outer strip.

Next, wrap a length of bus wire around the inner screw and tighten the screw into the wood. Stretch the bus wire so that it just touches center of the strip and hangs off the far edge. Hang a small weight, fishing wire or a few washers, off the end of the wire as close as possible to the shaft while making sure the weight does not the shaft itself. Trim any excess wire.

Do exactly the same thing to the other terminal.

Your rotating platform is now powered. If you find that you are having trouble getting the power through then use two bus wires from each screw.

Method Two

Another method (not my favorite for this application, but a good general purpose technique) to build a usable slip ring starts by going to Radio Shack, or another electronic supply house and purchasing a copper-clad board like the kind used to make printed circuit boards (PCBs). Radio Shack sells an etching kit for hobbyists who want to make their own PCBs. You don't need to buy the kit, just the board and some Ferris Chloride etching solution.

The copper clad board is usually sold in a rectangular shape. Lay the rectangle in front of you so the width is longer than the height and then, using a pencil, draw a vertical line dividing the rectangle into two equal halves. You are going to do exactly the same thing to each half to produce two identical boards.

Suppose we start on the left side. Measure the width of the head tube assembly which holds the bicycle wheel. Draw a circle in the center of the left side of the circuit board that is just a bit larger than the diameter of the head tube assembly. Now, draw four concentric circles, each radius being one quarter inch larger than the last. We'll call the inner-most of these four circles number 1 and the outer-most circle number 4. Using a black marker, fill in the space between the circles 1 and 2 leaving the space between circles 2 and 3 bare. Then fill in the space between 3 and 4.

Do exactly the same to the other half of the copper-clad board.

Now, pour the etching solution into a glass baking tray and submerge the board completely for at least half an hour. Carefully inspect it and make sure that all the copper has been etched away. If not, resubmerge it until it is. Then, using a little acetone, gently wash away the black ink to reveal the four copper rings.

Cut the board in half, separating the rings into two sets of two. Then cut out a center hole in each side that's just large enough to slip the head tube through. Trim the boards so they have the final appearance you want. You may want, for instance, to cut the corners off the rectangles and create squares or even circles. Make sure to leave at least one half inch of space between the outer-most ring and the edge of the board. Drill two screw holes on opposite sides of the boards so you can secure them to the clinostat.

You will only do this next step to one set of rings, so put the other in a safe place. You'll need some copper wire that's braided, that is, wire that is built up from many smaller wires twisted together. Gently strip this wire completely into two 4 foot strips. Now, every eighth of an inch or so along each ring, drill a small hole all the way through the copper and the board. The hole should be just large enough to insert the stripped wire. Next, ring each set of holes with a length of bus wire to form a continuous circle bar wire on the backside of the board (the side without the copper rings). Hold the bus wire in place with a few tabs of tape.

Next, you're going to stitch the braided wire through the holes almost as if you were sewing a shirt. Start from the back (the side with no copper rings) of the inner ring. Insert and pull the wire through. Now, insert the back through the next hole (this time from the copper-ring side) and pull it almost all the way through, but leave a loop of wire on the copper-ring side that's about one quarter inch long. On the backside of the board, pull the wire across the bus wire as you insert the braided wire into the next hole. You want the braided wire to make contact with the bus wire between each hole. Continue this process pulling the wire flat to the bare side (across the bus wire) and leaving quarter inch loops on the copper ring side until you've completed the ring. Now, using a separate piece of wire, do the same thing for the outer ring.

Using a hot soldering iron, solder each side of each loop to the copper right where the wire exits the hole. Be careful not to let the solder rise up the copper wire. Also, solder it everywhere it crosses the bus wire on the back side of the board. The wire should be firmly connected to the copper at each hole and also to the bus wire at every junction.

Next, trim the loops with a pair of wire cutters until to 1/8 if an inch from the board. Spread out the braids. You should now have a large number of small brushes electrically connected to the copper rings and to the bus wire on the back.

Next, take the board without the braids and drill two small holes between the copper rings. These holes will be used to connect the copper rings to the motor.

The two parts of the assembly must be pushed together to keep the contact. I use a 1 foot long flexible, clear plastic ruler to provide the restoring force. For now, cut one in half and duct tape the cut edges to opposite sides of the non-braided assembly. We'll get to using these as springs in a few steps.

Take the two boards and position them so that they are facing each other. Slip them together as a single assembly over the head tube and then slip the head tube through the rest of the supports. The board with the wire brushes should be flush against the vertical support while the other board rests freely on the head tube. Solder the power leads to the bus wires, one going to ground and the other going to the positive 12 volt supply. Connect the wires from motor driving the wheel to rings on the second disk by passing the wires through the holes drilled in the non-braided disk and soldering them into place.

Push the two plates together so the brushes make a firm contact with the non-braided rings. Then flex the halves of the ruler inward and duct tape them to the shaft so that they provide a spring force which holds the braids against the unbraided copper rings. Make sure you don't make the connection so strong that the braids scratch the copper rings. If they do scratch, the copper rings will be quickly warn away and your slip ring assembly will be useless.