Introduction to the Problem

McGraw Tower has been home to the now famous pumpkin for over 5 months. Its longevity has made many question its authenticity. It seemed that a real pumpkin would have fallen off after a few weeks of exposure to the harsh Ithaca environment. Our goal was to devise a way to find out exactly what the object was made of.

The four of us had seen the publicity for this contest sponsored by the provost's office in the Daily Sun and heard the discussion around campus. We knew that this contest would prove to be a good challenge, but one that as a group we could overcome. We have been friends for some time and felt that this would provide us with a chance to try something new. All of the problems we had worked on in our classes were theoretical and we were never able to apply ourselves to solving a problem of this sort. With much enthusiasm we entered the contest.

In attacking this problem, we modeled our techniques on those used by astronomers in planetary exploration. Those techniques include:

1) Remote Sensing- The analysis of a target from a distance.
2) Flybys- Taking of data by a probe.
3) Sample Retrieval- The collection of a sample from a target.

Remote Sensing:

For remote sensing, we attempted to take CCD images through a telescope to examine the spectrum emitted by the object on the tower. Unfortunately, there was too much ambient light in daytime to do this successfully. The CCD was flooded by bright afternoon light, making the object nearly impossible to distinguish from a bright background. Furthermore, evening lighting was so poor that the object was indistinguishable from a dark background. This would have worked well if pumpkins glowed at night, but alas, this is not the case.

Flybys and Sample Retrieval:

The apparatus we designed on the night of Thursday, February 26, met both of these objectives. It consisted of a camcorder and hollow drill bit mounted on a platform, airlifted by weather balloon to the object. We built the apparatus over the course of the following four days and launched it early Wednesday morning, March 4. We were successful on our first attempt, obtaining both a sample and close proximity video footage of the object on the tower.

It is clear now from our analysis of the sample, as well as the video footage, that the object on the tower is indeed a pumpkin.

Construction and Control of the Apparatus

In developing the apparatus, we decided that we wanted a mechanism that we could control from the ground, something that could actually retrieve a sample of the object. We also wanted a way of getting close up video footage of the object. Mounting a camcorder on the apparatus and connecting it to a monitor on the ground not only allowed us to get the close up footage, but also gave us an excellent view to help guide the sample retrieval. In addition, by using the parts of a remote control plane we were able to control the moving parts of the apparatus from the ground.

Already having the camcorder and remote control components, we used the tools in Professor C.H.K. Williamson's engineering laboratory to design and build the rest of the apparatus. We purchased two sheets of Plexiglas( to serve as the base for the apparatus. To actually retrieve a sample, we bought a coring drill bit and some PVC piping upon which the bit was mounted. We also bought metal sheets out of which we cut hinges, allowing the drill to extend and retract.

We stacked the Plexiglas( pieces on top of one another, with the ball bearings from a lazy Susan between them. The bottom piece of Plexiglas( served as the base of the apparatus and was anchored to the balloon. The top piece served as the mount for the equipment that we used. We cut holes in the middle of the plates in which we placed a servo from the remote control plane. This allowed the top piece of Plexiglas( to rotate with respect to the base and balloon. This made it much easier for us to position the drill bit directly in front of the object once the apparatus was anchored to the tower spire. This anchoring was accomplished via the inclusion of two wooden dowels, each of which extended at an angle to the base. A piece of string stretched from one dowel end to the other, about three feet in front of the base. We were then able to encircle the spire, thereby anchoring the apparatus to the tower. This greatly improved our stability in the face of gusty winds.

The camcorder was bolted to the center of the mount, directly above the servo in the lazy Susan. It was cushioned by a piece of foam padding which prevented it from jarring against the mount. This central position gave it an excellent look at the target in the center of the field of view with the drill arm at the bottom. The drill bit at the end of the drill arm was mounted on a motor, which in turn was contained by the PVC piping. The piping rested on the hinges, and the hinges were bolted to the mount. The motor was controlled remotely by a servo attached a to paper clip. When rotated, the paper clip touched a wire, completing the circuit and turning on the motor.

One of these hinges on which the piping sat was connected to a servo. This servo could rotate the hinge, which in turn extended and retracted the drill arm. This made drilling the object on the tower much easier, as we could extend the arm to reach the object instead of waiting for the wind to blow the drill bit in close.

The apparatus was lifted by a helium-filled weather balloon eight feet in diameter, which had 18 pounds of lift. Our device weighed approximately 12 pounds, so we had about six pounds net lift. Clothesline extended upward from all four corners of the base of the apparatus, and met at a point four feet above the mount. We tied a piece of wire around the clothesline at this point, which extended upward another foot to the balloon.

Control of the apparatus was provided by three kite strings, one attached to the rear center of the apparatus base, and one each at opposite corners of the front. Each kite string was 500 feet long, far more than enough to guide our device to the top of the 173-foot tall tower. Using three strings made flying the apparatus a stable procedure. Using just one string would restrict the elevation of the device, but would give no control over side to side or forward to backward motion, as with a kite. Two strings, one at the front and one at the back, would restrict forward to backward motion but would do nothing for side to side motion. Three strings, one on the back and one on either side in front, gave us control over elevation, side to side motion, and forward to backward motion. The three lengths of string were wrapped around cylindrical reels, which when turned either let out or reeled in string. This is how we controlled the position of the apparatus.

Launch Day

The day chosen for balloon launch was Wednesday, March 4. We met in the parking lot by Uris Library at 5:30am. Winds were calm, and it was lightly snowing. We determined that these weather conditions were suitable for flight. We then assigned duties for preparation--Jon and Fred filled the balloon with helium while Sam and Eldar prepared the electrical equipment. The monitor and VCR sat in the back of Eldar's car to protect them from weather. Stereo wire was connected from the monitor to the camcorder to set up a live video feed. Once the balloon was filled with helium, it was attached to the apparatus and we were ready for launch. We launched the balloon from the south side of McGraw Tower at 7am.

We let the balloon rise higher than the tower spire, and then Jon, Fred, and Sam guided the balloon directly over the top of the tower by taking up positions at 120 degree angles around the base. Using binoculars and the live video feed, Eldar directed the others in fine-tuning the horizontal and vertical positioning of the apparatus. We were able to anchor the apparatus onto the tower spire at about 7:20am. Eldar then used the remote controls to maneuver the drill arm and retrieve a sample. Gusty winds at the top of the tower made this very difficult, but a small sample was retrieved at about 7:55am. At this time, Eldar directed the other three to guide the apparatus up off the spire and back to the ground. The apparatus touched down at about 8:10am, making for a total flight time of an hour and ten minutes. Once the apparatus reached the ground, the drill bit, which held the sample, was wrapped in cellophane and removed.

Ascending the Tower

Final Approach

Preparing for Sample Retrieval

Analysis:

As a first step, we wanted to determine whether or not the sample was a plant. To accomplish this, we examined the sample with a light microscope and a scanning electron microscope. We found cell walls and xylem vessels, and when exposed to ultraviolet light the sample displayed autoflorescence. These are all distinctive characteristics of plants. Most plants exhibit these traits, so this does not identify the object as a pumpkin. To determine specifically what plant we had we conducted DNA testing on our sample. We used two different methods of DNA testing.

The first method involved extracting DNA from the sample, duplicating it using a process called Polymerase Chain Reaction (PCR), then analyzing it by cross checking our DNA with a computer DNA database for over 3,000 known flowering plants.

The second method also entailed extracting DNA and using PCR, but in this case we duplicated the DNA with a special primer that only pumpkins and very few other plants respond to. We could then further test for pumpkin DNA by measuring the length of the strands and comparing it to the lengths of a known pumpkin DNA.

The PCR process began with placing our sample in a buffer solution and mashing it with a plunger. This broke the cell walls allowing DNA to flow out into the buffer solution. A primer was then added to the solution which allowed us to choose where the duplication would start. For the first method, we used primers that duplicated a sequence called RBCL, a long sequence in the chloroplast gene common to many plants. For the second method, we used primers that duplicated a sequence called MP27, a seed maturation protein. Polymerase was also added to the solution, which provided the proteins that would compose our duplicated DNA. Polymerase consists of the bases A, C, T, and G which are the building blocks of DNA. Enzymes were also added to act as catalysts to the reaction.

This solution was heated to 92(C to denature the DNA, separating the double helix into single strands. Then the solution was cooled to 50(C for the annealing phase. This was the time when the primers adhere to specific sites on the single strands. Finally the solution was heated to 72(C for the synthesis phase. In this phase, the polymerase starts binding to the end of the primer and continues down the length of the single DNA strand, faithfully replicating the strand's compliment. This process is repeated many times, effectively doubling the amount of DNA in each cycle.

To check what DNA was duplicated and how much yield was obtained, a sample was placed in an agar gel matrix and a voltage difference was applied across it. DNA is charged, so the potential gradient induces a force on it, causing it to migrate forward through the gel. The larger DNA moves slower, so the size of DNA can be determined by how far it has moved relative to a sample of known length. A dye is added to the DNA, allowing it to be seen under ultraviolet light, so the intensity of light gives a measure of the yield obtained.

For the first method of DNA analysis, we were allowed to use the laboratory of Jane Doyle and were assisted by her graduate students. We duplicated the RBCL strands and sent them to the Bio Resource Center for sequencing which gave us the exact composition of the RBCL. This data was then compared with 3,000 other plant species to determine its most probable identity and its nearest genetic relatives.

For the second method, we were assisted by Peter Kipp at the Boyce-Thompson Institute and used primers provided by the DNA Sequencing Facility. The MP27 sequence was chosen for duplication because it is a DNA strand in pumpkins that has very little similarity to DNA strands in other plants. Using these primers, the probability of reproducing a random DNA sequence accidentally was approximately one in a trillion. So if DNA was in fact produced, it was highly likely to be pumpkin DNA. To confirm this, we would add enzymes that would cut the DNA in specific locations. The length of the cut portions could be measured with the gel technique, so that we could confirm that the DNA strands were of lengths typical of a pumpkin.

The results of the analysis of the first method would hopefully show what family of the object came from, hopefully the pumpkin family or at least squash family. Since the RBCL sequence similar to sequences in other plants, it is still possible that the object could be one of the a closely related plant(. But based on its physical appearance, this is highly doubtful.

The results of the analysis of the first method are inconclusive. This is due to the widespread presence of RBCL in many other plant species. Unfortunately, this sequence of DNA was not unique enough to pumpkins to serve as an indicator of the presence of pumpkin DNA. DNA was definitely present, however, showing that there was something organic on the tower.

The results of the second DNA test were also inconclusive. We ran the PCR process for a known pumpkin, two tower object samples, and several other miscellaneous plants such as tobacco and a bacteria sample of E-coli. What we saw in the test was that the tower samples and the pumpkin sample behaved exactly the same, showing no bands at all, while the other plants showed several. The fact that these were missing in the pumpkin and tower object samples indicate that the primers were used up somehow. It is possible that this was due to having too much DNA in the buffer solution prior to PCR. We did not have enough time to attempt to correct this problem. It is significant to note that, while we did not obtain conclusive results, the tower samples behaved identically to the known pumpkin sample and not like any of the other plants that were tested.

Optical Image of McGraw Tower Object at 85x

These tube shaped cells are xylem vessels. They carry water and minerals up from the roots.

Optical Image of McGraw Tower Object at 98x

Dehydrated parenchyma cells. Large thin walled cells, very typical in plants.

Optical Image of McGraw Tower Object at 40x

Xylem vessels with iron contaminant (circle at right).

Optical Image of Known Pumpkin at 40x

Xylem vessels. Notice similarity with tower object.

Optical Image of McGraw Tower Object at 58x

This picture shows autoflorescence of lignin. Lignin is a hard material embedded in the cellulose matrix of the cell walls. It gives plants a "woody" structure. When ultraviolet light is shown on the cells, the lignin emits light at a different frequency. This was observed using Filter Florescence Isothiocyanate (FITC).

Scanning Electron Microscope of McGraw Tower Object

Dried out and broken xylem vessels with a parenchyma cell visible in the lower edge.

Scanning Electron Microscope of Known Pumpkin

Xylem vessels covered by layer of tissue. Similar structure as McGraw Tower Object, but not dehydrated.

Conclusions

Our analysis indicates that the object on top of McGraw Tower truly is a pumpkin. The close up images from the video footage show that the object is very similar to a pumpkin in size and color. The exhaustive analysis of the sample we retrieved shows that it definitively a plant, and shares many traits with a known pumpkin. So there is evidence on both the macroscopic and microscopic level to support our theory.

One argument disputing the authenticity of the pumpkin suggests that no real pumpkin could possibly last five months without decaying to the point that it would fall from the tower. Our video suggests how the pumpkin could have lasted so long. Shots of the top and bottom of the pumpkin show that a large circular hole was cut out of the center. Whoever placed it up there cored it out first, perhaps to make the placement easier. This coring would significantly reduce the amount of water in the pumpkin, and the large hole must provide good ventilation for air to flow through the center. These factors would combine to dry out the gourd in a short amount of time. Instead of rotting, the pumpkin has turned into a leathery husk, which could cling to the spire for decades.

Video shots of the drill colliding with the pumpkin also support this theory. The drill had difficulty penetrating the skin of the pumpkin; the small tears that the drill made indicate that the skin was brittle and very dry. Most of the matter that the drill managed to remove fell away. A moister specimen would probably have stuck fast to the bit.

In the end, we cling to the belief that the McGraw Tower Object is a pumpkin as strongly as the pumpkin has clung to the spire in the past months.

Acknowledgements

Professor Charles H. K. Williamson- For use of lab and equipment.
Don Cipolla- For relaying commands on launch day.
Mrs. Cipolla- For known pumpkin sample.
Mary Sansalone- For pictures of launch day and video stills.
Dr. Parthasarathy- For all microscope pictures.
Drell children- For use of walkie-talkies.
Professor Earl Kirkland- For use of pressure regulator for helium tank.
Professor Jane Doyle- For use of her lab.
Donovan Bailey, Chris Hardy, Carl Lewis- For guidance in DNA testing.
Michael Casaus- For pizza.
Bio Resource Center (Tom Stelick, Tatyana Pyntikova, Jennifer Griswold)-
Sequencing and analysis assistance.
Peter Kipp- For guidance in DNA testing.
DNA Sequencing Facility (Bob Sherwood)- For primers.