Marshmallow Catapult 2017: Pushing the Envelope

Since 1974, schoolkids across the USA have competed with each other in "Science Olympiads." The half-day struggles are divided into age groups (elementary, middle, and high school). An Olympiad consists of a set program of typically 16-19 separate events, ranging from pencil-and-paper quizzes to hands-on activities like analyzing unknown powders or ID'ing rock samples. I have coached in four Olympiads locally. My favorite events are the build type, where students design and build a device ahead of time and bring it to the competition to test against similar devices built by their peers at other schools. Because of my long standing interest in throwing engines, the elementary school Science Olympiad contest "Marshmallow Catapult" has become my favorite. This year (2017), our team won first place with the catapult described in the following Instructable.

I considered whether or not to keep our research and design process to ourselves, but that's not in the spirit of STEM education or the Science Olympiad. I decided to share complete details of our design and our prototypes, hopefully as an aid or inspiration to other teams.

Let me note here I am not connected to Science Olympiad in any way, other than as a volunteer coach. The opinions in this article are mine alone, and do not reflect the views Science Olympiad, my public school district, or anyone but myself.

For a history of Science Olympiad, see the article here.

The rules for Marshmallow Catapult are copyrighted, so I will paraphrase them. Important rules are italicized. Violation of these rule results in penalties to the team. Points I think are worthy of notice when designing the device are in bold.

A team of 1-2 kids have the task of building a small catapult out of certain materials. The purpose of the catapult is to launch a standard-size grocery store marshmallow at a target flat on the floor.

There are two targets, one close and one far. Each team gets one shot at each target. The team has to announce in advance which spot they are targeting. The targets can be anywhere from two to six meters away from the launch position, at half-meter increments. Relevant distances therefore are:

2 meters; 2.5 m; 3 m; 3.5 m; 4 m; 4.5 m; 5 m; 5.5 m; 6 m.

The exact target points for the day's competition are chosen by the event judges when the contest begins. The catapulteers are allowed five minutes to set up their device and take their shot.

The catapult must remain at the indicated start line. You can't move it around.

All members of the catapult team must wear safety glasses. Safety first!

Each team's machine is impounded at the beginning of the Olympiad and remains secured until the actual contest.

Design Requirements:

Power can be supplied by a single mouse trap spring or any number of rubber bands. Rat traps are specifically prohibited.

The catapult must be fired using a pencil to actuate a trigger or switch. Another safety idea--keep the kids' finger clear of moving parts.

In 2017, the catapult must fit in a square 'footprint' of 75 x 75 centimeters (roughly 29 x 29 inches). This is a fairly generous size. No restriction is placed on the height of the device or on the total weight. This rule may change from year to year, so check the official rules carefully.

The front of the device must be on the firing line during operation. A catapult should have some method of adjusting for the differing distances of the target. No specific requirement here; just that the device should be capable to throwing varying distances. The catapulteers are allowed to refer to recorded firing data they have accumulated in practice.

Not a rule but a guideline: most catapults are built of wood. Some builders use PVC tubing. Metal frames are not prohibited, and metal fasteners (screws, nails, bolts, nuts) are commonly used.

No metal is allowed on the launch arm except at the hinge. Another safety issue. Science Olympiad doesn't want metal parts flying off during the competition.

Every catapult must have something non-marring on the base, to protect floors from being damaged. These events are often held in school gyms, and coaches are very protective of their floors.

No practice launches are allowed. Two targets, two shots are all you get. Judges note the point of impact by the marshmallows and measure how far from the center of the target they hit. Closest to the bull's eye wins. The two shots are averaged together, so one good shot can be spoiled by a second poor shot. Misfires count as shots. In case of a misfire, the distance to the target is measure from where the catapult sits.

Catapults must be free standing. Teams are not allowed to hold or steady the devices when fired.

No part of the catapult can come off when fired. No pointed projectiles, liquids, flammable substances (!) are allowed. Pity! A flaming marshmallow would be very dramatic, and it would stick where it hit rather than bounce.

Violations of the rules are ranked in 3 levels. "Tier 1" means no violations. "Tier 2" violations occur if the team breaks one of the build or firing rules. "Tier 3" sanctions happen if the team removes their safety glasses, or if the judges decide the catapult is unsafe to fire.

Most competitors build a variation of the typical tabletop catapult, with a single throwing arm mounted in a frame, powered by rubber bands or a mouse trap spring. Some lightweight designs rely on the innate springiness of thin lengths of wood (like paint stirrers or tongue depressors). There's nothing wrong with this approach, and all of these designs can certainly toss a Jet-Puffed the required distance. Problem is, the concept is stale and the machines inaccurate. Let's throw off the the shackles of conventional thinking and start fresh!

The basic one arm, tossing style catapult derives from common ancient designs known variously as onagers, mangonels, lithobolos, tormenta, monagkones, and so on. All of these work the same way--a stiff throwing arm is pulled back against the tension of a power source (usually a skein of twisted cord, or at Science Olympiad, rubber bands). The projectile is placed in a cup or spoon at the end of the throwing arm. When released, the tension is released, throwing the arm forward in a long arc. The catapult frame stops the movement of the arm, while the missile--the marshmallow--continues on its way to the target.

This style of catapult can only throw things in a high arc, and it's very hard to hit a precise spot flat on the floor by tossing a light weight, irregularly shaped projectile like a marshmallow.

The first idea our team had when contemplating this problem was, "Why toss when you can shoot?" The boys on the team thought it wiser to propel a missile in a straight line at the target instead of tossing up in the air. A flat trajectory projectile can be aimed with more precision than one tossed in an arc.

It was an excellent idea. Several problems with the event would vanish if we could shoot instead of toss. The complicated process of trial and error, compiling data on how far a tossing machine would hurl a marshmallow at a given pullback, was no longer necessary. Because a shooting device can be aimed at the target, differences in propulsive tension became unimportant. With enough rubber bands, our machine would fling a marshmallow with equal vigor at any target, near or far.

Instead of an onager, our machine would more resemble a ballista (or in modern terms, a slingshot). Ancient ballistas were two armed throwing machines, using two propulsive skeins set sideways to form a sort giant crossbow. Greek and Roman ballistas threw giant javelins, but the largest ones hurled rocks too.

Our first prototype was pretty much a mini ballista or crossbow, with a rigid T shaped frame with rubber bands attached at the ends of the T. This did not work well at all. It was hard to get the rubber bands to hold the marshmallow securely (it often just fell out before release--a misfire), Even when we added a pouch to hold the marshmallow, the missile very often flew wildly, hit the frame of the catapult, or just flopped at the team's feet.

We did some research and came across something called a stonebow. This was a hunting weapon widely used in the Middle Ages and after until displaced by blackpowder firearms. A stonebow is a crossbow specially adapted to hurl balls or bullets. Usually the stock of the stone bow is curved out of the way of the projectile (see picture). The stone bow uses a double bowstring too--a useful clue for the final form of our catapult.

Even with a stonebow-style doubled bowstring, marshmallows kept slipping out and falling to the ground with embarrassing regularity. Clearly something more secure was needed.

One of the boys on the team remembered seeing an ad for a device called the Pocket Shot (TM). This is a handheld slingshot that uses a ring-shaped frame and a piece of cone-shaped rubber for power. All you have to do is drop a projectile into the rubber cone, hold the ring, drawn back the shot inside the cone, and let fly. Centering the shot in the cone causes it to fly through the tubular holder without impact, and at high velocity. Now we were on to something!

We could not simply adapt the Pocket Shot design to launch marshmallows. The manufactured rubber cones are too small. and the Pocket Shot system requires the missile to be pinched inside the cone during the draw. This is OK for hard plastic pellets or ball bearings, but a marshmallow would by mashed by such treatment. What evolved from the Pocket Shot concept was a tubular yoke made from a 4 inch diameter PVC plumbing coupling.

A 4 inch PVC coupling is a very sturdy item. It would easily stand up to any reasonable amount of rubber bands we could attach to it. Since two strands had failed to hold our marshmallow missile in the stonebow prototype, we tried four. This worked well, but we went on to try five and then eight strands. This was too much power. The catapult almost tipped over under recoil, and it was hard for the catapult team (two boys, aged 10 and 11) to cock the device against the pull of eight strands.

(At this point we were using two No. 33 sized rubber band looped together as one strand. Four strands was therefore 4 x 2, or eight No. 33s total; five strands were 5 x 2, and eight strands 8 x 2.)

The 2017 catapult was built of pine, poplar, and oak dimensional lumber purchased from a local home center. Keeping in mind the entire device had to fit inside a 75 x 75 cm footprint, here are the dimensions of the catapult:

height of stand, 36 inches

length of stock, 24 inches

length of the base, 26 inches

length of the base cross beam, 24 inches

width of the stand, 3 inches

width of the stock 1.5 inches

Dimensional lumber in the USA is only nominally measured. The ubiquitous "2x4" stud hasn't measured 2 inches by 4 in years. The lumber listed below is given in actual size. In home centers they use nominal sizes--thus the poplar studs I show as being 1.5 inches square are called "2x2," the oak planks are called "1x3" and so forth.

Materials for the stock:

(width x length x thickness)

(2) 4 x 24 x 0.5 inch yellow pine planks

(1) 2.5 x 21 x 0.75 inch oak plank

(1) 2.5 x 6.625 x 0.75 inch oak plank

(1) 1.5 x 17 x 0.75 inch oak

(2) 1 x 0.25 x 4 inch poplar lathe

(6) 8-32 bolts, 2 inches long each

(8) 8-32 nuts

(14) flat washers to fit 8-32 bolts

(4) small 0.5 inch long wood screws, round head

(1) 10-24 hex head bolt, 3.5 inches long

(2) flat fender washers

(1) 10-24 wing nut

Materials for the pedestal stand:

(width x thickness x length)

(2) 1.5 x 0.75 x 36 inch poplar

(1) 1.5 x 0.75 x 28 inch oak

(8) short deck screws, about 1.5 inches long each

(1) 1/4-20 hex head bolt, 4 inches long

(2) flat fender washers

(1) 1/4-20 wing nut

Materials for the base:

(1) 1.5 x 1.5 x 26 inch poplar

(1) 1.5 x 1.5 x 24 inch poplar

(4) 1.5 x 1.5 x 9 inch poplar

(8) short deck screws, 1.5 inches long

Self adhesive non-slip rubber pads

Materials for the launcher & trigger:

(1) 4 inch diameter PVC pipe coupling

(1) 3 x 3 inch 90 degree angle steel bracket

(1) 2 x 2 inch 90 degree angle steel bracket

(1) 0.75 inch diameter neoprene washer with a 0.5 inch hole

(1) 0.75 inch long 10-24 round head screw

(1) 10-24 nut, with lock washer

(1) 2 inch long 10-24 bolt

(2) 10-24 nuts

(1) screen door hook. The mounting eyelet should be removed; you only need the hook.

(2) No. 33 rubber bands

(4) Brites File Bands (7 x 1/8 inch rubber bands)

(5) 3mm white nylon zip ties

(1) 0.75 inch diameter faucet washer with a 0.25 inch hole

about 18 inches of steel pipe hangar strap

Tools: power miter saw, drill press, hand drill, screwdriver, small adjustable wrench, protractor, marker pen, super glue, assorted drill bits, pencil.

Materials for the stock:
(width x length x thickness)

(2) 4 x 24 x 0.5 inch yellow pine planks

(1) 2.5 x 21 x 0.75 inch oak plank

(3) 8-32 bolts, 2 inches long each

(3) 8-32 nuts

(6) flat washers to fit 8-32 bolts

The catapult stock is simple lamination of three layers of wood. The outer layers consist of two pieces of 4 inch wide, half inch thick yellow pine planks, cut to 2 foot lengths (24 inches). Choose which edges are to be on top and bevel 3.75 inches of the top edge. Use a 22.5 degree bevel. This is the only cut I could not do at home; I went to a local carpentry shop, and the gentleman there made the cuts for me in short order on his massive table saw. I squared up the ends of the bevels with an X-acto saw. The high end of the bevel should be on the outside. This creates a 'cup' for the PVC coupling to rest in. You can fill in the void under the yoke with a 4 inch length of 0.75 x 0.75 square rod--pine, poplar, or oak doesn't matter. This will give strength to the anchoring bolts later when the metal strapping is added.

Cut the 2.5 x 0.75 oak plank to a length of 21 inches. This is the hard core of the stock. Fit it between the 24 pine planks so that the front and bottom edges are flush. Clamp them together and drill four 1/8 inch holes along their length, avoiding the center point at 12 inches. Put a 2 inch long 8-32 bolt through each hole, after first put a suitable washer under the bolt head. Put another washer on the open end, then an 8-32 nut. Tighten firmly.

Cut a 2.5 x 0.75 inch oak plank to 6 and 5/8 inches. This will be the trigger mount, set at right angles to the stock and filling the last 3 inches of the pine planks. Study the photos. Set the trigger mount in place and drill for its 8-32 bolt, washers, and nuts as before, but don't install it yet!

In the next step we'll make the catapult's yoke.

Materials for the yoke:
(1) 4 inch diameter PVC pipe coupling

(4) Brites File Bands (7 x 1/8 inch rubber bands)

(4) 3mm white nylon zip ties

about 18 inches of steel pipe hangar strap

(2) 8-32 bolts, 2 inches long

(2) 8-32 nuts

(4) flat washers to fit 8-32 bolts

(1) 1.5 x 17 x 0.75 inch oak

The yoke, where the rubber bands are attached, is a simple but stout piece of PVC plastic. A four inch diameter coupling is more than strong enough to handle any forces generated by the catapult, and the diameter allows marshmallows to pass through with ease.

First you will need to mark out the holes needed to attach the rubber bands. Set the coupling vertically on a flat surface. Center a common protractor on the open end of the coupling and mark the positions at 0, 90, and 180 degrees. (I used a black Sharpie permanent marker). Put dots on the coupling's rim at the indicated places. Rotate the protractor 180 degrees and set the 0 and 180 marks on the protractor on the previously made marks. Then place a dot at the 90 degree position. This will give you 4 dots on the rim of the coupling 90 degrees apart.

With a small square, transfer the marks about half an inch down the side of the coupling. Drill a 1/8 hole through the PVC at each point.

Insert a 3mm nylon zip tie through the first hole. Put the zip tie through the loop of one of the Brites file bands (or whatever rubber band strand you are using). Close the zip tie and tighten down, but leave enough space in the tie loop for the rubber band to move forward and back without binding. Repeat this procedure on the other three holes until you have four strands of rubber bands attached to the coupling.

A few notes about the rubber: at first we used a pair of No. 33 rubber bands looped together to make one strand. These worked fine, but the bands wore out quickly and broke during practice. Also, when fired, the bands hurtle forward, driving the marshmallow through the yoke toward the target. As a result, the bands tend to tangle. The bands must be untangled before shooting again. Shooting with tangled rubber results in wild shots or misfires. Always untangle the band before firing again!

Late in the build we discovered Brites file bands. These are brightly colored synthetic bands measuring 7 inches long (unstretched) by 1/8 inch thickness. Four of them do the work of 8 No. 33s, and they last a lot longer. Also, the long bands tend to tangle less than the built-up strands. Brites come in color sets, too. By color coding the bands, it makes the task of keeping the bands untangled easier. We used two blue bands on top and two orange bands for the lower pair. A quick glance at these will disclose whether the bands were in their proper orientation.

The yoke should be mounted with the mounting holes forward. This maximizes draw length, putting more force into the projectile. When mounting the coupling to the catapult stock, the four holes (and rubber bands) should be set at 1:30/4:30/7:30/10:30 clock positions. Take care to align the coupling carefully. Accuracy depends on precision.

The yoke will nestle in the beveled cut at the front end of the stock. It's held in place with a length of steel pipe hanging strap, about 18 inches worth. Wrap the strapping snugly around the center of the coupling. Drill a hole through the stock near the bottom end of the strap and put a 2 inch 8-32 steel bolt through. Use washers under the bolt head and on the open end, then add an 8-32 nut. Tighten firmly. About two inches above this first bolt drill a second hole and add another 8-32 bolt, washers, and nut. When you tighten this bolt, it will draw the strapping tight around the coupling. This is a very secure mounting method. If needed, you can add two short wood screws through holes in the strapping in the upper half of the coupling. These will anchor the strap firmly to the yoke.

Finally, insert a piece of 1.5 x 0.75 x 17 inch oak in the groove behind the yoke. It should be a good friction fit. This piece of wood will further stiffen the stock and act as a stop against the yoke.

(1) 2.5 x 6.625 x 0.75 inch oak plank

(2) 1 x 0.25 x 4 inch poplar lathe

(4) small 0.5 inch long wood screws, round head

(1) 8-32 bolt, 2 inches long

(3) 8-32 nuts

(4) flat washers to fit 8-32 bolts

(2) No. 33 rubber bands

(1) 3 x 3 inch 90 degree angle steel bracket

(1) 2 x 2 inch 90 degree angle steel bracket

(1) 2 inch long 10-24 bolt

(2) 10-24 nuts

(1) 0.75 inch diameter neoprene washer with a 0.5 inch hole

(1) 0.75 inch long 10-24 round head screw

(1) 10-24 nut, with lock washer

(1) screen door hook. The hook's eyelet is what you need; you don't need the hook itself.

(1) 3mm nylon zip tie (or 1 1-inch steel split ring key ring--see text!)

With the yoke installed and Brite bands in place, it's time to make the trigger system. Science Olympiad rules require a trigger that can be tripped with a pencil.

Our first trigger was simply a hole drilled vertically in an oak block, set at a right angle at the end of the stock. A 6-penny nail shoved loosely in the hole was the trigger. To shoot, the bands were drawn back on their keeper and ring and hooked over the nail. Yanking the nail out fired the catapult. This worked, but it was crude. Pulling the nail up and out also disrupted the ring and keeper, spoiling accuracy.

(A word about the ring and keeper--we gathered the four rubber bands together in this order: top right upper band, bottom right lower band, bottom left lower band, top left upper band. These were looped in order on a piece of twine. The ends of the twine were fed through the center hole of a 0.75 inch black neoprene washer. This washer is the "keeper," as it keeps the bands together when shooting. It is the keeper that actually strikes the marshmallow and drives it out of the device. Behind the keeper we used a split spring steel ring--a 1 inch diameter key ring, actually--threading the rubber band strands through the split and rotating the ring until the ring was captive on the bands. This arrangement functioned very well, but it is not legal under 2017 Science Olympiad rules! Because the rules say "no metal on the throwing arm except at the hinge," we had to replace the split ring with a 3mm nylon zip tie. Use a good quality zip tie, as the strain of firing may break a cheap one. We competed with a nylon zip tie and won first place, but the photos show a steel split ring.)

Wanting a more exact mechanism that a nail and a hole, we next tried using a clothespin as a spring loaded trigger. We drilled a hole on one side of the clothespin, near the jaws, and fitted a smooth steel pin. You could then draw back the rubber bands and secure the anchor ring in the jaws of the clothespin, where the steel pin held the ring in place until the open end of the clothespin was pressed. Zing! This worked well, but again accuracy suffered, as the rising nose of the clothespin carried the ring with it, throwing off the flight path of the marshmallow.

What was need was a trigger than didn't disturb the anchor ring so much. We then cut the anchor block taller, settling on a 2.5 x 6.625 x 0.75 inch block. It was essential the rubber bands were centered when the mechanism was cocked, so we made a centering jig. I got a 4 inch PVC pipe 'knockout' cap. This fit inside the yoke, of course. We drilled a 1/8 inch hole dead center in the cap. With the cap in place in the yoke, we shone a laser pointer through the hole. Where the red dot touched the trigger block we cut a horizontal slot about 7/8 inch deep and 1/8 inch wide. Then we drilled a 1/4 inch hole through the top of the block down into the slot for a pivoting pin to slide up and down in.

After much discussion and fiddling using bent nails, wooden levers, and so on, we settled on a 2 inch right angle steel bracket. Using pliers, we bent the bracket from an L shape to a flattened V, sort of like this:

---v---

W fixed short pieces (0.25 x 4 inch) poplar lathe on either side of the trigger block to keep the bent bracket in place. 4 short wood screws hold the lathe in place. Otherwise the trigger lever floats; it's not fastened to the block at all, but pivots on the bend in the 'V'.

For the trigger pin we used a screen door hook with the hook cut off (we used a Dremel tool and cut-off wheel; a hacksaw would do as well). The resulting pin had a ring at the top and a straight shank about an inch long. We tried this simple lever first, but the mechanical advantage was low, so we bolted a 3 inch right angle bracket to the rear of the bent V trigger. To prevent the trigger coming out of its place if pressed too hard, we attached a 3 inch long 8-32 bolt to the last open hole on the 3 inch bracket to act as a stop, to prevent over-travel. Study the photos--they explain it best.

To keep the trigger normally closed, we looped together a pair of No. 33 rubber bands and threaded them through the eyelet on the trigger pin. We added extra nuts and washers to the bolt we used to fix the trigger block to the stock. We put a washer, nut and second washer tight to the bolt head before inserting it through the stock. a second 'sandwich' of washer, nut, washer nut was applied to outside end and snugged in place. This arrangement gave us a handy place to hook the rubber bands tensioning the trigger.

Worried the screw joining the two brackets might work loose, we put a drop or two of super glue on the nut. Make sure the screw slot is aligned 90 degrees across the axis of the catapult. We used a slotted screw for this connection for a planned reason: the final sighting system.

Nothing in this project caused more experimentation and prototyping than the aiming system. It was a essential aspect of our design philosophy that our catapult must be aimable, that we shoot at the target, not just toss a marshmallow toward it. When the design of the yoke, stock, and trigger mechanism was finalized, we had to tackle the question of how to aim. We tried:

1. Mounting a modified laser pointer in a hollowed out cork, loaded in place of a marshmallow. When the laser dot settled on target, the cork was removed and standard marshmallow loaded.

2. Mounting a laser pointer in a flat nylon washer the same diameter as a marshmallow and inserting it like a marshmallow, as above.

3. Using a protractor and length of string to align the catapult on the target and determine the optimum angle of aim.

4. Legend marks on the rim of the yoke were used to align on the target spot be peeking down the stock, like aiming a rifle.

5. Mounting a laser pointer on an offset mounting below the yoke and 'dialing in' the point of aim.

6. Mounting a cheap air rifle telescopic sight on a rail connecting the trigger block to the yoke.

None of these were satisfactory, though (4) led to the final design.

The problem with using a laser is that marshmallows are wonky projectiles. They're light, often misshapen, and vary in weight from one to another. Over longer distances they tumble and drift through the air. The result of these variables was to make exact sighting methods unprofitable in terms of time and effort. Science Olympiad sets a time limit of five minutes for the event. The team has to set up, load, aim, and fire at both targets within five minutes. There's no time to adjust a telescopic sight or play with plumb lines and protractors.

Using black marker dots on the rim of the yoke gave surprisingly good results, and the method was fast. All the team had to do was sight down the stock and center the target in the yoke. The legend marks made this easier.

Originally the marks were remnants of the drilling process (marking off 90 degree segments where the rubber bands could be attached). With just three legend marks (zero, ninety, and 180 degrees) we were able to zero in fairly well on the target. Improving on the idea, we cut pointed indicators out of wooden tongue depressors and screwed these into place where the black marks had been. (See photos)

The final touch came in the form of a rear sight. A flat steel washer was super glued in place in the slotted screw head on top of the trigger bar. This made an excellent "ghost ring" sight. Now all the team had to do was peer through the rear sight, line up the target with the front indicators, and let fly.

Side note: aiming shots were not allowed. Too bad! Given 3 shots, we could consistently zero in and shoot actual bull's eyes at least on closer targets.

Materials for the pedestal:

(width x thickness x length)
(2) 1.5 x 0.75 x 36 inch poplar

(1) 1.5 x 0.75 x 28 inch oak

(8) short deck screws, about 1.5 inches long each

(1) 1/4-20 hex head bolt, 4 inches long

(2) flat fender washers

(1) 1/4-20 wing nut

Materials for the base:

(1) 1.5 x 1.5 x 26 inch poplar

(1) 1.5 x 1.5 x 24 inch poplar

(4) 1.5 x 1.5 x 9 inch poplar

(8) short deck screws, 1.5 inches long

Adhesive rubber non-slip pads

Accuracy requires a steady platform, and we experimented with a few designs before settling on a straightforward pedestal model.

Our first catapult mount was an old 35 mm camera tripod. This appealed because it allowed aiming with its pivoting traverse and elevation controls. Unfortunately, the 75 x 75 cm footprint meant we couldn't extend the tripod's legs to their widest and most stable extent. With a narrowed base, the light aluminum tripod was not steady with the somewhat heavy catapult mounted on it. When fired, it rocked, so we had to abandon the camera tripod. Under other circumstances, the tripod might have been an excellent choice.

With the tripod idea discarded, it was back to wood. Because the Science Olympiad rules did not limit the height of a catapult, we could make our pedestal as tall as we wanted. It would make our task more sure if we could shoot down at our targets. The only practical consideration was the height of our catapulteers. In the end we settled on a 36 inch pedestal as a good compromise between ease of aiming and accuracy.

The central pedestal is an I-beam, made of a 1.5 x 0.75 x 28 inch oak center strake faced by two 1.5 x 0.75 x 36 inch poplar pieces. Leave a 2.5 inch gap in the oak at the bottom to fit over the base, and several inches of space at the top to allow the stock room to pivot up or down. (See photos). We used four short deck screws per side (8 total) to bind the pieces together, staggering them for the strongest grip.

Our first base was made of common 2 x 4 lumber, but we had trouble finding lumber that wasn't warped. We gave up on 2 x 4s and resorted to "2 x 2" poplar, which are actually 1.5 inches square. Poplar is excellent straight-grained wood, clear and strong. Keeping in mind the required footprint parameters, we made the base of two interlocking pieces of "2 x 2" poplar. The piece parallel to the stock is 26 inches long; the crossing piece is 24 inches. I cut a half-lap joint in the base pieces just ahead of the center of the 26 inch piece. We wanted the pedestal centered on the bottom rail and the cross piece just in front of it. We put a short deck screw through the intersection, then fitted the pedestal in place. Carefully we bored a 1/4 inch hole through the pedestal outer planks and through the base, making very sure the pedestal was flush with the underside of the base. A 4 inch 1/4-20 bolt with one large fender washer under the head and another on the open end secured the pedestal to the base. We put on a 1/4-20 wingnut, thinking this would allow us to dismantle the stand if needed.

The bracing turned out too complex. I had four pieces of "2 x 2" poplar cut at 45 degrees on each end, salvage from an earlier project. Because of the hollow center of the I-beam and the offset base, I had to cut small filler blocks from "1 x 2" poplar to make everything fit. This could have been done more simply.

To prevent marring any floors (per the rules) we applied stick-on rubber pads to the underside of the base. These also help keep the catapult steady when fired.

The last thing to do was mount the complete catapult on the stand. I had a 10-24 bolt, washers, and wingnut handy, so we bored through the stand and stock and bolted them together. (I actually used a drill press for this, as it was important the hole be at 90 degrees to the stock; otherwise the catapult mounting would be askew). With the wingnut snugged in place, the catapult was held in a secure friction fit. Traverse was limited to nudging the stand left or right, and elevation was best done by the boys using a fist to bump the stock up or down.

Was this all too much for an elementary school science contest?

Not at all. We worked on the catapult from early January to mid April 2017, meeting for about an hour once a week after school, with time out for Spring Break and other holidays. As the adult, I did all the sawing and drilling, but the boys brainstormed and tested each new idea we came up with. We disagreed at times; for example the boys were very attached to using the camera tripod as a stand, but its instability eventually became too obvious. We tried very hard to use a laser pointer in some way (lasers are cool!), but in the end, the waywardness of marshmallows showed us a laser was unnecessary.

We learned a lot. We learned about strength of materials, elastics, aerodynamics, physics, the flight path of irregular objects, and how to work within a set of rules while still being creative. Our design was radical enough I anticipated other teams might object to it, so I secured approval from the head of Science Olympiad in our state. It was a good thing I did, as some coach at the event did protest that our machine was not a true catapult. Every other machine in the contest was a onager type. Fortunately we had prior approval for our design.

And we won first place! The real purpose of Science Olympiad is not to win medals or trophies, but to take part; to use the competition to experience the scientific method firsthand. We were given a task to perform. Problems developed, theories were tested, solutions were found. The events are meant to be fun and thought-provoking. Marshmallow Catapult certainly succeeded in making the team and I think, and I believe we all had fun doing it.

Books we consulted while researching Marshmallow Catapult:

GURSTELLE, William. Backyard Ballistics. Chicago Review Press, 2012.

MIDDLETON, Richard. The Practical Guide to Man-Powered Weapons and Ammunition. Skyhorse Publishing, 2007.

NOSSOV, Konstantin. Ancient and Medieval Siege Weapons. The Lyons Press, 2005.

PAYNE-GALLWEY, Sir Ralph. The Crossbow. Barnes and Noble Books, 1996.

(Many editions of this book are available.)

RIHILL, Tracey. The Catapult: A History. Westholme Publishing, 2007.

TUNIS, Edwin. Weapons, A Pictorial History. World Publishing, 1972.

(Modern editions of this title can be found.)

A fun and informative website on all things slingshot: The Slingshot Channel.