Engineering and Science

C a l i f o r n i a

I n s t i t u t e o f

T e c h n o l o g y engineeringE& andS science

Volume LXXI, Number 1, 8

i n t h i s i s s u e

The Final Frontier

The Last Big Slice of Pie

The Fates of Cells On March 11, 2008, space shuttle Endeavor carried not one, but two Caltech alums to the Interna- tional Space Station. Here, astronauts Garrett Reisman (MS ’92, PhD ’97), left, and Robert Behnken (MS ’93, PhD ’97) posed for the camera before Behnken embarked on the mission’s fourth spacewalk. After the successful mission, which included installation of a module of the Japanese laboratory, Kibo, on the ISS, Endeavor touched down on the night of March 26 at Cape Canaveral—but without Reisman. He will remain on the ISS as the flight engineer, not returning until space shuttle Discov- ery, scheduled for launch on May 25, comes to pick him up. Reisman also took some special cargo on behalf of his former advisor, Hayman Professor of Mechanical Engineering Chris Brennen, whose wife suddenly died of cancer in August 2007. Accompany- ing Reisman on his cosmic journey are the Bren- nens’ wedding rings, fused together. engineering& and science California Institute

o f T e c h n o l o g y Volume LXXI, Number 1, 2008 E S 2 Random Walk

12 50 Years in Space — by Douglas L. Smith International leaders in the space and aeronautics industry convened at Caltech to share their thoughts on the first 50 years of space exploration—and what’s in store for the future.

22 Auctioning off the FCC’s Crown Jewels — by Elisabeth Nadin When the FCC recently auctioned off bandwidth previously reserved for broadcast television, Caltech researchers were enlisted to ensure that everyone got a fair shot at the prized goods.

28 When Cells of a Feather Don’t Flock Together — by Marcus Woo Cells don’t behave the same way, even if they’re nearly identical. So how do they decide what to do?

34 Ad Astra per GALCIT — by Douglas L. Smith A new master’s degree program launches graduate students into the space industry.

39 Letters

Obituaries: Seymour Benzer, David C. Elliot, Herbert B. Keller On the cover: Mission spe- 41 cialist Robert Behnken (MS ’93, PhD ’97) seen during a six-hour, 53-minute Engineering & Science (ISSN 0013-7812) is published Robert M. Kieckhefer spacewalk, his first, outside quarterly at the California Institute of Technology, 1200 President of the Alumni Association the International Space East California Boulevard, Pasadena, CA 91125. Annual Robert L. O’Rourke subscription $20.00 domestic, $30.00 foreign air mail; Vice President for Public Relations Station. Behnken and fel- single copies $5.00. Send subscriptions to Caltech 1-71, low mission specialist Rick Pasadena, CA 91125. Third-class postage paid at Pasadena, Linnehan were installing CA. All rights reserved. Reproduction of material con- STAFF: Editor — Douglas Smith tained herein forbidden without authorization. © 2008, Writer — Marcus Woo parts for a two-armed California Institute of Technology. Published by Caltech Contributing Writers — Signe Bray, Elisabeth Nadin, Jill Canadian servicing robot and the Alumni Association. Telephone: 626-395-3630. Perry, Jackie Scahill, Kathy Svitil, Michael Torrice Copy Editors — Allison Benter, Barbara DiPalma, Michael named Dextre. PICTURE CREDITS: Cover, inside front cover, 13, 17, 19, 20, 38 — NASA; 2 — Paul Rothemund, Robert Lang; 3 — USPS; 8, Farquhar, Elena Rudnev 12 — NASA/JPL-Caltech/U. of Arizona; 11 — Audrey Wu, Dick Business Manager — Debbie Bradbury Sneary Illustrations; 13 — USGS, NASA; 18 — Boeing, NASA/ Circulation Manager — Susan Lee SVS; 19 — NASA/JAXA, NASA/JPL; 20 — NASA/GSFC; 15, Photographer — Robert Paz 21, 36, 37 — Bob Paz; 22, 26 — FCC; 25 — Jacob Goeree; 28-33 Graphic Artist, Layout — Douglas Cummings — Elowitz lab; 23, 25, 31 — Doug Cummings; 34–35 — NASA Landsat; 37 — Caltech Archives; 42 — Elliot family; 44 — ACM Visit us online at http://EandS.caltech.edu Random Walk

F ROM THE L AB TO THE G ALLERY

Right: Rothemund’s DNA origami of North and South America ren- Matisse, Picasso—and tions of the AFM scans etched dered as a three-dimensional glass now, DNA and computa- into glass blocks, using the etching. tional origami. Science, art, same techniques used to make technology, and design come laser-etched glass paper- Below: Lang’s Snack Time depicts together in a new exhibit at weights. the wedding feast of a female the Museum of Modern Art Lang has combined his love praying mantis on her unfortunate in New York. The show, en- of mathematics and paper- male partner. And yes, this was titled Design and the Elastic folding, becoming one of the Mind, includes work by two pioneers in computational folded from a single uncut square Caltech alums—origamist origami—the art and science of paper. Robert Lang (BS ’82, PhD of designing origami with ’86) and Paul Rothemund (BS mathematical techniques. ’94), a senior research associ- The exhibit showcases some of ate in computation and neural his intricate creations, as well systems and computer science. as the TreeMaker software he Rothemund, one of wrote to make his increasingly Caltech’s two 2007 MacAr- complex designs. thur “genius” grant winners, Both types of origami are invented “DNA origami,” displayed alongside a myr- in which he turns strands of iad of provocative exhibits, DNA into any desired ﬂat including a honeycomb vase, shape, from a smiley face to pig bone tissue grown into the the outline of a contient. He shape of wings, and furniture took DNA from a harmless modeled after human bones. virus and developed a method The show runs through May to fold and pinch strands 12. —MW together. The result is a powerful way to build nanoscale structures out of DNA. The shapes measure about 100 nanometers, or 100 billionths of a meter, across—about one thousandth of a hair’s breadth. In addition to atomic-force micrograph (AFM) prints of Rothemund’s creations, the exhibit includes representa-

Right: Have a nano day!

2 ENGINEERING & SCIENCE NO . 1 2 0 0 8 On March 6, the post office issued its second series of four American Scientists stamps. This set includes Caltech Nobelist Linus Pauling (PhD ’25). The background art refers to his discovery that sickle-cell anemia is a molecular disease—the first to be recognized as such. Caltech is two-for-two in this series of stamps: the first set, issued in 2004, included fellow laureate Richard Feynman.

G ETTING N ANOWIRED

One day, they could be on the other hand, is one of from larger systems like car is not only cheaper, but can everywhere, powering your the most abundant elements engines, and may also be used also be done with lower- computer and keeping its mi- in the universe. The micro- in refrigeration devices. The quality silicon. The trick to croprocessor cool at the same processor industry has also researchers, who include Wil- turning nanowires into solar time. They’re silicon nanow- made processing silicon inex- liam Goddard (PhD ’65), the cells is a unique geometry, an ires, narrow devices hundreds pensive and easy. But because Ferkel Professor of Chemistry, idea first developed by gradu- to thousands of times thinner silicon is also an excellent con- Materials Science, and Ap- ate student Brendan Kayes than this piece of paper. Two ductor of heat, it didn’t seem plied Physics; Jamil Tahir- (MS ’04) in 2005. Regular groups of Caltech researchers promising as a thermoelectric Kheli (MS ’86, PhD ’92), a solar cells are flat, and absorb are discovering the remarkable material—until now. senior staff scientist with the photons face-on. The newly properties of silicon nanow- By growing silicon into Materials and Process Simu- freed electrons then move ires, enabling the wires to nanowires, researchers in lation Center; and gradu- along the same direction, par- harness solar power and to act Heath’s lab improved silicon’s ate students Akram Boukai allel to the incoming photons. as refrigerators by converting thermoelectric efficiency by (PhD ’08), Yuri Bunimovich They’re collected at the surface heat to electricity, and vice a factor of 100. One of the (PhD ’07), and Jen-Kan Yu, of the silicon slab, where they versa. reasons for the enhanced per- reported their findings in the then join the electrical cur- The latter group, led by formance might be a phenom- January 10 issue of Nature. rent. Additionally, the cells James Heath, the Gilloon enon called phonon drag, ac- Silicon nanowires may also have to be thick enough to Professor and professor of cording to the team. Phonons help solve the energy crisis. capture all of the photons. chemistry, found that silicon are heat-carrying vibrations Researchers in the labs of In the new photovoltaic could be an efficient thermo- that travel across the material. Nate Lewis (BS, MS ’77), the cells, silicon nanowires sit electric material when made Constricted by the small size Argyros Professor and profes- alongside one another like into wires only 10 nanometers of the nanowire, the phonons sor of chemistry, and Harry blades of grass. Light is (10 billionths of a meter) don’t scatter off the sidewalls Atwater, the Hughes Professor absorbed along the length of wide. “At these tiny dimen- in the nanowire. Instead, they and professor of applied phys- the wires, which, at tens of sions, nature is doing things travel unimpeded down the ics and materials science, are microns, are still long enough that were previously not wire and drag electrons with using the wires to build a new to snatch all the photons. The thought possible,” he says. them, which improves ther- kind of photovoltaic cell. advantage of this configura- In a thermoelectric mate- moelectric performance. Most conventional cells are tion, however, is that the rial, a difference in tempera- Although the silicon made from silicon wafers. In- electrons move widthwise— ture sends electrons scurrying nanowires are still only about coming photons from the sun perpendicular to the photons’ to the cooler end, creating a half as efficient as state-of-the- are absorbed by the silicon paths. The nanowires are only current. To be efficient, the art thermoelectric materials, and dislodge electrons from several microns in width, so material must conduct elec- further improvements—as their atoms. The electrons are the electrons don’t have to tricity well; but to maintain well as lower manufactur- then free to move, producing travel as far, allowing them a temperature difference, it ing costs—could make these enough current to power cal- to produce electricity more must conduct heat poorly. tiny devices useful in a host culators, light bulbs, and even easily. Once the electrons are Most thermoelectric materi- of applications. They can entire homes. The drawback collected in the outer shell of als efficient enough to be make microprocessor chips is that these solar cells must the wire, they quickly travel to useful are expensive and hard more efficient by recovering use pure, top-quality silicon, the top of the wire and enter to make, restricting them to leaked heat. Eventually, they which is expensive to process. the current. niche applications. Silicon, may be able to recover heat Growing silicon nanowires The team has made nano-

2 0 0 8 ENGINEERING & SCIENCE NO . 1 3 wire arrays one square centi- meter in area—orders of mag- D ANCES WITH D N A nitude larger than any made before. The researchers have also been able to embed the nanowires in a flexible membrane for added versatility. Nature is a software the molecules take is actu- locked up in the inaccessible The membrane is excellent at engineer par excellence. By ally the goal of our programs, output domains—when the absorbing light, as is evident rearranging protein and and the destination is just the output domain pairs up with from its near-black color. RNA building blocks, nature by-product: it’s what you get still other pieces of DNA, the The best conventional programs myriad molecules when the function is com- entire system goes to a lower- silicon solar cells are about 25 to synthesize, haul, detect, plete,” says Pierce. energy state. percent efficient at convert- and regulate one another. The dancers are short “hair- By designing how each ing sunlight to energy, says Now scientists are trying their pins” of DNA that fold back hairpin domain pairs with its postdoc Michael Filler. The hands at it, and in the January onto themselves. Each hair- fellows, the Caltech team can researchers’ nanowire cells are 17 issue of Nature, a group of pin has three domains—one harness this energy to make just over one percent electri- Caltech researchers pub- input and two output—that the molecules perform the cally efficient. But Filler lished examples of their own can interact with domains in specific task they want. This says they are making rapid molecular programs. Associ- other hairpins by matching part of the design process is progress, and are aiming for ate Professor of Applied and the “letters” in one strand the most difficult, and re- 20 percent efficiency. “Our Computational Mathematics with the letters in another. In quires the team to model the group has been pushing the and Bioengineering Niles the alphabet of DNA, A pairs physics of the hairpins. Using forefront of the field right Pierce, senior postdoctoral with T and G matches C; the algorithms developed by now,” he says. Other mem- scholar Peng Yin, grad student hairpins contain between 50 Pierce’s group, the researchers bers include postdocs Ste- Harry Choi, and research and 100 letters. The hairpin’s ensure that when the hairpins phen Maldonado and Kate technician Colby Calvert input domain is initially are mixed together, they in- Plass, and graduate students showed how molecules of available to pair up with other teract appropriately so that no Michael Kelzenberg (MS ’06), DNA only ten nanometers in DNA molecules, while its hairpin runs off with another James Maiolo, Leslie O’Leary, length could be directed to output domains are inacces- hairpin’s dance partner. Morgan Putnam, and Josh perform specific tasks unaid- sible. Once a matching piece To showcase the hairpins’ Spurgeon (MS ’06). Once ed—without external energy of DNA binds to the input capabilities, Pierce and his col- the researchers achieve higher sources, temperature changes, domain, the hairpin pops leagues “wrote” four different efficiencies, Filler hopes in- or enzymes. open and the output domains programs. In each case, the dustry will jump in and push Biomolecular engineers are exposed. Output domains hairpins were designed not the design toward commercial have assembled DNA mol- of open hairpins can then to interact until an initiator use within the next decade. ecules into stable patterns, seek out the input domains molecule was introduced to —KS/MW like planar crystals, wireframe of closed hairpins and open the system. For the first pro- cages, tubes, smiley faces, those molecules. The ensuing gram, the initiator triggered and maps of North America. cycle becomes a molecular the hairpins to self-assemble Pierce and his colleagues square dance with hairpins via a specified sequence of concentrate on the motion of exchanging partners according “handshakes” into branched the interacting DNA mol- to the design of the bioengi- structures with multiple ecules. To see the distinction neers. arms shooting out from a between these approaches, All of these exchanges occur central point, like three- and consider the difference without Pierce having to add four-armed starfish. Upon between a choreographer energy to the system. So what completing this assembly (Pierce) and a cheerleading makes them go? “The basic process, the initiator then dis- coach. The coach primarily feature of the hairpin is that assembled from the structure cares where the cheerleaders it’s initially trapped in a high- to catalyze the formation of end up in a human pyramid: energy state,” says Pierce. This more starfish. In the second stronger, heavier people go on state is similar to a mouse- program, the hairpins assem- the bottom, while more agile, trap that has been set and bled into a tree-like pattern lighter people are at the top. baited. Until a mouse trips called a dendrimer, growing He doesn’t care how they get the trigger, the trap is stable from the root of the tree to there: all’s well that ends well. and doesn’t move. But within the leaves. Another program By contrast, when directing the spring of the trap, there is demonstrated a phenomenon dancers, the choreographer energy waiting to be released called autocatalysis, in which cares most about how the when the unsuspecting mouse a chemical reaction—in dancers move across the floor goes for the cheese. A piece this case, the production of and who they partner with. of DNA binding to the input fluorescent pairs of hair- Where they end up is of less domain triggers the hairpin pins—feeds on itself. After importance. “The trajectory to release the stored energy the initiator was added to the

4 ENGINEERING & SCIENCE NO . 1 2 0 0 8 solution, the test tube would low bioengineers to write in a program—for example, into a specific set of DNA begin to glow, getting brighter molecular programs the way an output domain on strand hairpin sequences to be syn- and brighter exponentially. that computer scientists write C binding to an input domain thesized. “We want to liberate The most dramatic example electronic ones. A compiler on strand D in step four, the molecular engineer from was a DNA “walker” that used is the software that translates only to unbind again later in having to think about the de- its DNA “legs” to lurch along high-level programs written preparation for the next cycle. tailed structural features of the a DNA track one step at a in a language like C++ into These reaction graphs are not molecules and instead focus time. The walker was inspired the binary instructions the unlike the flow charts beloved on the functional behavior of by the protein kinesin, which machine actually executes. As of computer programmers. the system,” says Pierce. glides along protein microtu- a first step, the team has devel- As a software package, the “In designing a compiler, bules in cells to move molecu- oped graphical representations molecular compiler would there’s work for many differ- lar freight. “Years ago, I was of their hairpins that are used translate an engineer’s design ent fields: computer science, amazed when introduced to in schematics, called reaction ideas into reaction graphs and applied mathematics, control the programmable chemistry graphs, to describe each step then translate those graphs and dynamical systems, chem- of kinesin. I decided then and there that I wanted to be able to engineer that kind of Atomic force micrograph by Peng Yin; graphics by Colby Calvert. molecular complexity. We still have a long way to go,” says Pierce. Left: How to grow a ninja-star Pierce foresees these hairpins being put to use as dendrimer. The Matrix-esque molecular sensors or nano- glyph to the left of each stage is mechanical drugs. Molecular the reaction graph, or program- instrumentation could detect small changes within cells, ming instructions, for making it. like the switching on of a gene Left, inset: The key to the reac- within a developing embryo, tion graphs. Each circular “node” producing a fluorescent signal represents a DNA hairpin. Input for scientists to read and analyze. He hopes that pro- domains are triangles; output grams like the exponentially domains are circles. Inacces- glowing one could develop sible domains are filled with solid into cheap technologies that colors. Binding to an accessible would amplify the presence of a miniscule amount of an domain will open the inaccessible interesting molecule into a ones. Arrows between the nodes detectable signal. “Instead indicate what binds to what. of thinking of instrumentation as something expensive Below: The dendrimer is built that your experiment resides from five kinds of A hairpins and within, we want to design four kinds of B hairpins that exquisite instruments that you assemble themselves in sequence embed within your system of study,” says Pierce. Program- according to the reaction graphs. mable molecules may also eventually lead to dynamic drugs that use one input domain to pinpoint cancer cells, triggering an output domain to kill them. For Pierce, the work represents a step toward the long-term goal of developing a compiler for biomolecular functions that would al- Inset and bottom figure reprinted by permission from Macmillan Publishers Ltd. from Yin, et al., Nature, vol. 451, January 17, 2008, pp. 318–323. Copyright 2008.

2 0 0 8 ENGINEERING & SCIENCE NO . 1 5 istry, and physics,” says Pierce, who is teaming up with other researchers at Caltech and elsewhere on the project. “Building a molecular compiler is a very daunting challenge, but progress in the field has been pretty dramatic in the last five years, so a primi- tive first-generation compiler is probably now within reach,” C ALTECH C ONNECTS WITH L OCAL C LASSROOMS says Pierce. Even so, the day when dancing molecules detect and kill cancer cells in humans is probably still far in As a Caltech grad student, I tion within teacher-Techer the future. —MT never get to work at a quarter pairings are pretty loose; each after 7 a.m., but when I join pair decides how to spend science teacher Tobias Jacoby their time together. For some in his classroom at Blair High teachers, science is a little School in Pasadena, this is outside their comfort zone, when my day begins. Mr. and they enjoy the confidence H OW D O Y OU G ET TO C ARNEGIE H ALL ? Jacoby and I have been paired boost of having an expert up through the Caltech Class- volunteer on hand to field dif- room Connection (CCC), an ficult questions. “It’s so great outreach program that brings to have someone with me Practice! And that’s exactly contrast to the other pieces, together Caltech graduate who I don’t have to explain what everyone in the Caltech- which are much louder,” ac- students, postdocs, faculty, everything to, because he just Occidental Concert Band has cording to Bing. and staff with Pasadena teach- knows,” gushed one teacher, been doing. On Saturday, Paul Asimow, associate ers. The hope is that putting describing her volunteer at a May 24, about 60 musi- professor of geology and people who practice and love recent CCC dinner. Other cians, including Caltech and geochemistry at Caltech, science into the classroom teachers—like Jacoby, who Occidental College students, will be conducting “Be Glad might inspire students to take is perfectly at ease explain- Caltech faculty and staff, JPL Then, America,” by William it more seriously—in their ing torque, momentum, and employees, and members of Schuman, which Asimow has course work, and maybe also kinetic energy—can really the local community, will known since playing as a stu- in their career plans. use an extra set of hands, as perform at Carnegie Hall. dent at Harvard. Asimow says The guidelines for interac- well as someone to bounce “We are extremely excited the piece treats the timpani as about this,” says senior physics melody makers, not rhythm major and clarinetist Lauren instruments. “Our timpanist, Porter, who has been integral Scott Babcock, is one of the to organizing the trip. “It’s a few professional members of huge opportunity for us, and the band, and I am happy to the culmination of a lot of give him this opportunity.” hard work.” Also featured is vocal soloist Band director and artist- Kjerstin Williams (BS ’00, in-residence William Bing, a MS ’02, PhD ’06), on George professional trumpet player, and Ira Gershwins’ “Someone has performed at such venues to Watch Over Me.” Wil- as Lincoln Center, the Ken- liams has been a trombonist nedy Center, and Disney with the Caltech jazz and con- Concert Hall, but this will be cert bands since her freshman his first apperance at Carn- year, but singing brings her an egie Hall. He handpicked indescribable thrill. “To sing the concert’s pieces to fit the with a wall of music behind venue. For instance, “Chorale you, there’s nothing quite like and Alleluia,” by Howard it,” she says. “Karaoke doesn’t Hanson, was chosen because even begin to touch it.” it suits the renowned acoustics If you live near the Big Thalia Reyes (left) and Maria Murillo, physics students at Pasadena’s John of the hall. “A Prairie Hymn,” Apple and would like to catch Muir High School, try to turn on a light-emitting diode (LED) with a bat- by Joseph Curiale, on the the show, visit http://www. tery made from a potato, a zinc nail, and a penny during a CCC-assisted lab other hand, was chosen for its carnegiehall.org or call 212- session. Photo by their physics teacher, Dave Herman. “meditative quality, and it’s a 247-7800 for tickets. —JS

6 ENGINEERING & SCIENCE NO . 1 2 0 0 8 ideas off of for new labs and I demonstrate the use of the me, as it has many, that Caltech’s Moore-Hufstedler activities. motion tracker for one group, science as taught in schools Fund, the Mattel Children’s During my classroom visits, sliding the cart slowly along doesn’t really tell kids any- Foundation, and the National I have mostly been helping the track as the position and thing about how scientists Science Foundation, secured out with labs and problem- velocity are plotted on a lap- do their jobs. In my class, I with the help of faculty solving sessions. Today I am top screen. “Wow,” said one hope to use some of the time director Christina Smolke, helping out with a physics girl. “I actually understood devoted to magnetism to tell assistant professor of chemical lab on collisions, and before that because someone actually students about one of the engineering. This generous the students arrive we set up explained it to me.” She just tools I use in my own neu- support has enabled the hiring five-foot-long, low-friction made my day. roscience research: magnetic of James Maloney (MS ’06) tracks on tables around the It’s a challenge to come up resonance imaging. Maybe as full-time codirector. He classroom. We place two with activities aligned with with the help of some cool acts as an ambassador for the brick-like carts with magnetic the state’s science standards, pictures, showing detailed project, visiting local schools front bumpers on each track. using inexpensive and readily brain structures and specific and sitting on education The students will roll one cart available materials. CCC regions that light up when committees. His presence has into the other, stationary cart volunteers are succeeding people learn reward asso- helped secure an ample supply to observe how momentum admirably and having fun in ciations, I can impress upon of teachers who are interested is conserved in elastic and the process. One volunteer them that physics is impor- in participating in the pro- inelastic collisions. In an used Kool-Aid to demonstrate tant for many fields of study: gram—what they need now is elastic collision, the magnets the concept of molarity to a biology, psychology, medicine, Caltech volunteers to match. repel each other, and the carts chemistry class. The students and engineering. (If you think you might be rebound. In an inelastic colli- made several batches of Kool- The CCC works with the interested in participating, sion, the bumperless ends col- Aid with different proportions Pasadena Unified School Dis- please visit http://www.class- lide with a satisfying “thunk,” of powder and water, and trict, which has seen its ups roomconnection.caltech.edu/ and the moving cart comes then calculated the concentra- and downs through the years. or email [email protected].) to rest, sending the stationary tion—or molarity—of sugar As a district where a high pro- Graduate students Tara cart rolling down the track. A in each, by assuming that portion of parents send their Gomez and Jennifer Franck motion detector positioned at Kool-Aid is 100 percent glu- kids to private schools, the (MS ’04) codirect with Malo- one end of each track records cose. They then related how public system is left to fend ney. The trio is extremely the positions and velocities of the drinks tasted to their sugar for itself. In my classroom, proactive about providing the carts over time. concentrations. Another the diverse group of students support to volunteers. “We There are 45 students in volunteer demonstrated that is friendly and open with each want to make sure that grad this class, one of two large energy could take on different other, and Blair strikes me as a students are getting some classes Jacoby teaches in addi- forms by powering an LED safe and genial place to learn. good teaching experience tion to a smaller IB (interna- with a battery made by stick- That said, class sizes are large, from the program, and we tional baccalaureate) group. ing a zinc nail and a copper resources are scarce, and state hope that this will help them At first glance they seem to be penny into a potato. Volun- test scores are often below decide if they want teaching masterfully combining high teers are endowed with a small standard. to be a part of their future,” social energy with academic budget for supplies; after Founded in 2002 by Eddie says Franck. lethargy: friends giggle and purchasing a graphing calcula- Branchaud (MS, PhD ’06), It’s really striking how much chat and seem to pay no mind tor for his class, a volunteer then a mechanical-engineer- fun volunteers are having with to the assignment at hand. demonstrated for his group ing graduate student, the students at all levels. As a They make the people I inter- how optimization can be used CCC has grown from just a scientist walking into a third- act with in an average day at to figure out the most efficient few teacher-Techer pairings grade class, “you get treated Caltech seem awfully sedate. combination of ingredients to to 20 this year. This growth like a rock star,” says Maloney. I circulate among the make Cheez Whiz. has been possible thanks to Kids that age are so naturally groups. There are some ques- As I relive the high-school funding from the Howard curious that a simple, hands- tions about the instructions. physics curriculum, it strikes Hughes Medical Institute, on demonstration can be the

2 0 0 8 ENGINEERING & SCIENCE NO . 1 7 basis for a great lesson. My high-school group is a little more aloof, and at first I worried that they would just dismiss me as a hopeless nerd. They very well may, but I’m not in high school anymore— M ARTIAN A VALANCHES so who cares? My students always make me laugh, and if I can say or do some small thing to help them squeeze a JPL’s Mars Reconnaissance Orbiter has captured the first-ever passing grade out of the tor- picture of Martian avalanches in action. The dramatic image ture session that is high-school below, taken on February 19, shows billowing clouds marking physics, then we will both be the course of dust and ice spilling down a 700-meter-high cliff happy. —SB that slopes at more than 60 degrees. The High Resolution Imaging Science Experiment (HiRISE) camera shot a swath of terrain some six kilometers wide by more Signe Bray is a graduate than 60 long near the Martian north pole. The frame captured student in computational and at least four avalanches—in this close-up of one of them, the dust neural systems who does brain cloud is 180 meters across. The white to the left in this false- imaging in the labs of Professor color rendering is carbon-dioxide frost at the top of the cliff. of Biology Shinsuke Shimojo This action snapshot provides a rare glimpse of the Red Planet’s and Assistant Professor of Psy- geology in motion. Scientists will compare it with previous shots chology John O’Doherty. of the area, and more observations through the Martian summer might reveal details about how the ice erodes. —MW

8 ENGINEERING & SCIENCE NO . 1 2 0 0 8 M ARKETING G ETS INTO Y OUR H EAD

Do you often lust after the sample five wines identified perceive more expensive items frontal cortex and another most expensive item on the only by their price. as higher quality. But does it region involved in process- shelf? Unbeknownst to the eager taste different, or do people ing emotional information, You’re not alone. After all, tipplers, however, two of rationalize? We didn’t know.” the amygdala, mirrored this expensive stuff is coveted—by the wines were the same but To answer this question, preference. definition. Otherwise, why labeled with two different the researchers looked at what But this study is the first to would people pay exorbitant prices, one markedly higher was going on in participants’ show that marketing actions, prices for things they really than the other. For example, brains while they sampled the in the form of hefty price don’t need? a $90 wine was presented wines. They used functional tags, can have an effect on the Take wine, for example. sometimes as a $10 wine and magnetic resonance imag- brain. The authors propose On surveying a wine menu in other times at its true retail ing, a technique that takes a that activity in the orbitofron- a fancy bistro, you might be price. three-dimensional snapshot of tal cortex reflects a value that tempted to judge the quality After tasting the wines, activity throughout the brain the brain assigns to the wine of the wines by their price. people were sometimes asked at a rate of about once every that combines information And why not? The more to evaluate either the intensity two seconds. They com- about its taste and its price. expensive wines are prob- of the flavor or the pleasant- pared brain responses to the Activity levels are higher the ably better, and will likely be ness of the taste. It turns out wines presented as expensive more impressive the overall a tastier accompaniment to that a $90 wine doesn’t taste to responses when the same value, teaching the brain to your tuna carpaccio. nearly as good when you wines were presented as less make this excellent choice Now suppose that your think that it costs $10. Both expensive, and found that the again. usually cheap date orders an wines that were presented medial orbitofrontal cortex The brain’s propensity to expensive bottle before you at two different prices were was more active when people integrate outside knowl- sit down, but you take a sip rated as more pleasant when tasted the more expensively edge into what we think are assuming that she chose her identified with the higher labeled wine. This region is internally generated opinions usual house red. Would you price tag. However, the flavor located above and between might make humans seem enjoy it more had you known intensity ratings, which acted your eyeballs, and is involved like dangerously manipulable that she made an uncharacter- as a control question, were not in processing experiences we creatures. But we evolved istic splurge? A new study led affected by the labeled price. deem rewarding, like winning in social groups, so why by Caltech Associate Professor Follow-up questions showed money and smelling food. not make use of the group’s of Economics Antonio Rangel that participants truly believed Activity in this area was cor- wisdom when making deci- (BS ’93) suggests that yes, the that they tasted five distinct related with people’s expressed sions? If you are unable to mere knowledge that a bottle wines. enjoyment of the wine, which ascertain the value of an item is pricey can cause you to Eight weeks after the tended to be greater the more for yourself, integrating other enjoy it more. initial study, participants expensive the bottle. people’s impressions into your In a paper published in were invited back to taste the This is not the first study judgment might not be a bad the January 2008 issue of the wines again, this time without to show that information idea. Proceedings of the National any price information. Not culled from sources other than Unfortunately, the wisdom Academy of Sciences, Rangel, surprisingly, without the price our noses and taste buds can of the group is not going to postdoc Hilke Plassmann, tags, the difference between influence our enjoyment of help you pay for that expen- Associate Professor of Psychol- two samples taken from the a smell or taste. An earlier sive bottle, or prevent you ogy John O’Doherty, and same bottle disappeared. And study by an Oxford University from indulging in regret- Stanford Professor of Market- this time, the wine people research team led by Edmund table trends. If your brain ing Baba Shiv describe per- liked the most was actually Rolls tested the impact of can trick you into thinking forming a little bit of trickery the cheapest—a $5 bottle. labels on our perception of an something tastes better than it on a batch of study partici- “In marketing, people odor. They gave participants a does, could this explain those pants recruited largely from spend a lot of money to whiff of cheddar cheese while terrible ’50s food fads? Spam- the Caltech community. “We create brand associations in a computer monitor displayed and-fruit-cocktail gelatinized advertised we’d pay people people’s minds, and establish either the words “body odor” party loaf, anyone? —SB money for tasting wine— a price-quality relation,” says or “cheddar cheese.” Not everybody was willing,” says Plassman, “and we know that surprisingly, people preferred Plassmann. During the study, it works. Marketing stud- the scent labeled as cheese. participants were asked to ies demonstrate that people Activity in both the orbito-

2 0 0 8 ENGINEERING & SCIENCE NO . 1 9 T HE P LANET H UNTER

with the solar system loved by together smoothly, as if, well, children everywhere would the eight planets were aligned. have to be changed. This For Rusch and Brown, dramatic revision would have the final product was a suc- to be explained to kids, says cess. According to Brown, Rausch, who has written for The Planet Hunter: The “The story is dead accurate,” both adults and children. “It including such details as his Story Behind What Hap- wasn’t so much that [Brown] failed childhood attempts at pened to Pluto discovered the 10th planet,” rocketry, illustrations of his by Elizabeth Rusch she says, “but that our under- childhood dog Roscoe and the standing of the solar system green 1964 Volkswagen Beetle Illustrated by Guy Francis was going to change one way he drove in high school, and Rising Moon, 2007 or another, and the story be- the discovery of Eris and the 32 pages, $15.95 hind that had to be told.” She demotion of Pluto. He’s not wanted to show that science sure how successful it’ll be in was about constant discovery bookstores, but “it’s fun to involving real people, and not watch the Amazon rankings,” just facts to be memorized he says. At the time of this from a textbook. writing, it’s number 138,554, She pitched the idea to her and at one point reached the editor, got quick approval, top 25 in the category of sci- and, just a week after Pluto ence and technology biog- High-achieving researchers don’t fit either category, how got the boot, sent Brown an e- raphies for children, joining often get their share of the do you get a children’s book mail requesting an interview. Einstein, da Vinci, and others spotlight. They win presti- written about you? Initially unsure about the who are old and dead. The gious prizes, go on national In January 2005, Brown request—admittedly an odd Planet Hunter has also been television and radio, have discovered an object in the one for an academic—Brown nominated for a 2008 Pacific books, and newspaper and Kuiper Belt, a population agreed to talk and help with Northwest Book Association magazine articles written of at least 70,000 icy bodies the book. Swamped with me- Award. about them and their dis- beyond the orbit of Neptune. dia requests and other distrac- The fact that Brown has a coveries. But rarely do you But this object, nicknamed tions from the “crazy Pluto young daughter helps him un- see a scientist featured in an Xena (as in the television thing,” as he calls it, Brown derstand the book’s audience, illustrated children’s book. series’ Warrior Princess), was then forgot about the project. he says. Lilah, who appears Professor of Planetary bigger and farther away than “The funniest part of this, in the story, loves the book. Astronomy Mike Brown now Pluto. Controversy swirled at least to me, is that I have “She thinks it’s the book has that distinction. He is the over whether the shiny sphere, almost no memory of this,” about her,” he says. hero of The Planet Hunter: The now officially named Eris, Brown says. “There were so The last illustration portrays Story Behind What Happened was indeed the 10th planet. many other things going on, Brown with Lilah a few years to Pluto (Rising Moon, 2007), If it wasn’t a planet, Pluto this was just one of a million older, looking at the night sky a children’s book written by shouldn’t be either, since both things that were happening.” with a telescope. “When I Elizabeth Rusch and illus- objects are markedly different In fact, Brown neglected to was flipping over the proofs, trated by Guy Francis. Rusch in size, orbit, and location— tell his wife until the book I saw that, and my heart just tells the story of Brown’s among other characteristics— came out in December 2007. sort of melted,” he says. “I childhood, his discovery from the other eight planets. Since the topic was a timely have a soft spot for the book of Eris—briefly known as Rusch had been engrossed one, the project had to move just for that.” —MW the 10th Planet—and the in the debate since the begin- swiftly. The fact that it took subsequent vote that demoted ning, she says. In August just over a year from pitch Pluto and made headlines 2006, the controversy came to publication is an anom- around the world. to a head as the International aly, Rusch says. Normally, A perusal of children’s Astronomical Union con- children’s books can take three books about scientists turns vened in Prague to vote on the years to publish. The edi- up stories about the likes of definition of a planet. Rusch tor found an illustrator, Guy Albert Einstein, Galileo, and realized the implications of Francis, who, as it turned out, Leonardo da Vinci—subjects the vote were huge, regard- illustrated the favorite book of who tend to be old and dead. less of the result. Every book, Brown’s two-year-old daugh- So if you’re like Brown and poster, and placemat adorned ter, Lilah. Everything came

10 ENGINEERING & SCIENCE NO . 1 2 0 0 8 A NNENBERG G ROUNDBREAKING

PICTURE CREDITS: 2 — Paul Rothemund, Robert Lang; 3 — USPS; 8, 12 — NASA/JPL-Caltech/U. of Arizona; 11 — Audrey Wu, Dick Sneary Illustrations Caltech will soon have a classrooms, an instructional new home for its interdisci- computer lab, and studio and plinary program in informa- office space for faculty and A ND S CHLINGER , TOO . tion science. The first institu- students. The center will also tion in the nation with such feature atrium and lounge a program, Caltech broke spaces to promote collabora- ground for the Walter and tion and interaction. The Leonore Annenberg Center building’s exterior is mostly for Information Science and glass, with a window in nearly Technology on December 7. every room, connecting the The field of information sci- structure with the campus, ence is as broad as it sounds, said Frederick Fisher, the encompassing many areas principal architect. of science and engineering The center will also herald from the theoretical founda- a new information-based tions of information to how curriculum at Caltech, and nature handles it in biological possibly beyond. “The dream systems to how it shapes social is very vivid in my mind,” systems. said Jehoshua “Shuki” Bruck, “When you’re crossing so the Gordon and Betty Moore many different disciplines, Professor of Computation and Chemists and chemical Katharine Schlinger. War- when you’re reinventing the Neural Systems and Electrical engineers at Caltech will soon ren (BS ’44, MS ’46, PhD very boundaries of science and Engineering and founding have a new playground. Work ’49) spent 12 years studying, the way it can improve our director of the Information on the Warren and Katha- researching, and teaching at lives, you deserve a research Science and Technology rine Schlinger Laboratory Caltech. Katharine grew up home, an intellectual cross- initiative. One day, he hopes, for Chemistry and Chemical in the Pasadena area and met roads that is as collaborative information will be taught Engineering began on Febru- her husband while working and inclusive and revolution- in schools and universities ary 13. as a department secretary for ary as the work itself. This alongside traditional subjects The lab will form a new Chemical Engineering. center will be that home,” like history or physics. focal point for the Division In addition to the Schling- said Wallis Annenberg, vice The Annenberg Center of Chemistry and Chemi- er and Moore Foundation president of the Annenberg aims for a Silver rating from cal Engineering, said David contributions, gifts have Foundation, which donated the Leadership in Energy Tirrell, the McCollum-Corco- come from an array of sup- $25 million to build the cen- and Environmental Design ran Professor and professor porters, including the estate ter. Stephen D. Bechtel Jr., a (LEED) Green Building Rat- of chemistry and chemical of former trustee Victor K. life trustee of Caltech, recently ing System, developed by the engineering. The new lab will Atkins; trustee G. Patricia awarded $1 million to the U.S. Green Building Council. attract new faculty and spur Beckman (daughter of Mabel project. Caltech hopes to Slated for completion in the research, said Tirrell, who also and Arnold Beckman, PhD raise a total of $31.5 million. summer of 2009, it will be chairs the division. ’28); Barbara J. Dickinson The 50,000-square-foot one of three new LEED-rated Located near the western (widow of Richard Dickinson building will contain an 80- buildings on campus. —JP/ end of the San Pasqual walk- ’52); the Ralph M. Parsons seat lecture hall, several small MW way on campus, the four-story Foundation; the John Stauffer building will occupy 60,000 Charitable Trust; John W. square feet and should be Jones (’41); Helen and Will finished in 18 months. It Webster (’49); Gregory P. will also likely merit a Silver Stone (’74); and others. rating under the Leadership Funds raised to date total $37 in Energy and Environmen- million; the building is an- tal Design (LEED) Green ticipated to cost $45 million. Building Rating System for —EN environmentally sustainable buildings. The Schlinger Lab was named in recognition of a $20 million campaign donation from Warren and

2 0 0 8 ENGINEERING & SCIENCE NO . 1 11 Fifty Years in Space

By Douglas L. Smith

An international Who’s Who of aerospace roughly the length of the auditorium, and luminaries packed Caltech’s Beckman Audi- Robert Goddard’s first rocket in 1926 torium last September to celebrate “50 Years reached about 41 feet, or approximately to in Space.” The conference was organized the ceiling. It took another 30-plus years by Ares Rosakis, the von Kármán Professor to reach orbit, with the Soviet Union’s of Aeronautics and Mechanical Engineer- October 4, 1957, launch of Sputnik ing and director of the Graduate Aero- as Earth’s first artificial satellite propel- nautical Laboratories at the California ling the world into the Space Age and Institute of Technology (GALCIT), and the United States into the Space Race. Dwight Streit, vice president, electronics Numerous speakers paid homage to the technology, for Northrop Grumman Space Cold War as a powerful motivator. Open- Technology, which cosponsored the event ing keynoter Ronald with GALCIT and the Jet Propulsion Labora- Sugar, chairman and tory, which Caltech administers for NASA. The CEO of Northrop heavy hitters from all of Earth’s spacefaring nations Grumman, said that were invited, says Rosakis, but the representatives “an international space from China and Russia were no-shows. According to movement based on Rosakis, the conference’s three chairs—Caltech presi- cooperation and the dent Jean-Lou Chameau, JPL director Charles Elachi quest for knowledge (MS ’69, PhD ’71), and Northrop Grumman Space emerged from one Technology president Alexis Livanos (BS ’70, MS ’73, originally based on PhD ’75)—represented the triumvirate of academia, geopolitical struggle national laboratories, and industry that has woven and military competi- the exploration and use of space into the fabric of tion.” Joanne Maguire, our society. The speakers, who were drawn from all executive vice president three branches of the triumvirate, celebrated past of Lockheed Martin accomplishments, reflected on our current situation, Space Systems, was and speculated on the future of humanity’s endeavors blunter. “One wonders in orbit and beyond. Herewith are some of their if the U.S. would have thoughts, culled from a day and a half of presentations done anything in space and panel discussions and reassembled in narrative in the 1960s if it were form. Streaming video of all the presentations, along not for the Cold War.” with event photos and speaker biographies, can be Infrared–NASA/JPL-Caltech Optical–NASA/STScI; X-ray–NASA/CXC/SAO; credit: Image DARPA, the found at http://galcit.caltech.edu/space50. Defense Advanced Supernova remnant Cassiopeia A through the eyes of three Research Projects Agency, was founded of NASA’s Great Observatories. The outer shell of cold dust WHERE WE’VE BEEN in 1958 along with seen in the infrared by the Spitzer Space Telescope is col- NASA, both in The journey into space began humbly enough, ored red, the filaments of warm gas visible to the Hubble response to Sputnik. as astronaut Ronald Sega, retired undersecretary Space Telescope are yellow, and the superheated shock wave That 84-kilogram of the U.S. Air Force, pointed out. The Wright sphere, clearly visible brothers’ first flight in 1903 was about 120 feet, seen by the Chandra X-ray Observatory is green and blue. every 98 minutes as

12 ENGINEERING & SCIENCE NO . 1 2 0 0 8 A false-color mosaic of it passed overhead, was a great shock to America’s Saturn’s far side from the presumption of superiority in all things scientific Cassini spacecraft. The and technical. David Whelan, chief scientist for Boeing Integrated Defense Systems and former rings are backlit, so the director of DARPA’s tactical technology office, said most opaque parts appear DARPA’s charter was to “prevent technological darkest.��Thermal radia- surprise,” and described how it functions as part tion from Saturn’s interior think tank and part piggy bank. DARPA foresaw the military usefulness of space for photoreconnais- lights up the night in red. sance, weather forecasting, telecommunications, Thick clouds deep in the and GPS. All of these, of course, have now found atmosphere block some civilian uses as well, and the synergy between military and civilian space efforts was a constant of it, appearing as dark theme of the conference. The agency started writ- Eugene Shoemaker points with his rock hammer as he streaks, spots, and globe- ing checks immediately, said Whelan, with ARPA describes the geology of Meteor Crater’s rim ejecta to encircling bands. Saturn’s (the word “Defense” wasn’t added to its name until astronaut trainees during a field trip in May 1967. sunlit side appears green- 1972) Order #1 going to Wernher von Braun to develop the first American spacecraft, the Explorer ish-yellow. 1. Later ARPA funding orders were for the devel- of the University of Texas had become Apollo’s opment of the Mercury and Gemini space capsules field-geology principal investigators. They planned that took Americans into orbit, and the Saturn mission-specific training trips, scouted appropriate rocket motors that took us to the moon. locations, and, perhaps most importantly, recruited On May 25, 1961, President Kennedy addressed mentors. These included, besides the three PIs, a joint session of Congress after the Russians had Caltech geology professors Lee Silver (PhD ’55) beaten the Americans once again by putting cos- and Robert Sharp (BS ’34, MS ’35), who “provided monaut Yuri Gagarin in orbit on April 12. Ken- infectious emphasis on ‘belly geology,’” and Rich- nedy challenged the nation to put a man on the ard Jahns (BS ’35, PhD ’43), a recent defector to moon by the end of the decade. NASA hopped to Stanford from the Caltech geology faculty by way it, and in 1962, Eugene Shoemaker (BS ’47, MS of Penn State. ’48) of the U.S. Geological Survey helped NASA Various speakers described the robotic post-Apol- develop a plan for lunar exploration. In January lo exploration of the solar system and our growing 1963 he began field-training astronauts, leading the understanding of the evolution of the cosmos and Gemini group—which included future moon- our place in it through a panoply of astronomical walkers Neil Armstrong, Pete Conrad, and John missions, including NASA’s Great Observatories. Young—on a two-day tour of Arizona’s Meteor This fleet of four space telescopes has covered the Crater and nearby volcanic features, said Apollo electromagnetic spectrum from ultra-energetic 17 astronaut Harrison “Jack” Schmitt (BS ’57). gamma rays down to the coldest of infrared radia- “Shoemaker was one of the best and most infec- tion. (The Compton Gamma Ray Observatory tious of teachers. I know that this trip impressed was de-orbited in 2000; its successor, the Gamma- my future colleagues.” By 1967, “science training reached a new NASA/Eugene Cernan NASA/Eugene level of sophistica- tion,” with spacesuited Harrison Schmitt, Apollo astronaut-geologists 17’s lunar-module pilot using the equipment and the second-to-last they would actually be carrying, operat- man to set foot on the ing under realistic moon, wields an adjustable time constraints, and sampling scoop. Apollo communicating their 17 brought back 109 observations by radio to scientists at mission kilograms of lunar material control. Shoemaker, from the Taurus-Littrow USGS colleague region in December 1972. Gordon Swann, and William Muehlberger (BS ’49, PhD ’54)

Three of NASA’s four Great Observatories are still in service.

2 0 0 8 ENGINEERING & SCIENCE NO . 1 13 ray Large Area Space Telescope, is slated for launch administrator, a congressman asked him what is a this May.) JPL’s Michael Werner, the project scien- very legitimate question if you don’t understand tist for the infrared Spitzer Space Telescope, called anything about space: ‘Why do we need weather the ensemble “a programmatic stroke of genius,” as satellites when we have the Weather Channel?’” they were built in sequence, maintaining a steady Ballhaus called space a utility. “It’s just like power stream of funding from year to year and from to your house. People don’t understand how the mission to mission, as opposed to the more usual power got there, but they sure notice if it doesn’t approach of starting with a clean slate annually and get there. . . . Last week we completely changed having to rejustify the budget. the ground system on GPS, and you didn’t read about it in the news, because it worked smoothly.” “The cost per year of Cassini, all of that three billion dollars amortized over Just like the water and electric companies, space utilities turn a profit. According to Boeing’s the life of the mission, is about what Americans spend annually on lip balm.” Whelan, “The U.S. government’s cost to pay for and manage GPS for the country is actually offset by the profit to the tax base due to all the revenues This assortment of intellectual riches led Neil from people buying cars with GPS units and so deGrasse Tyson, director of New York City’s forth. So it’s actually revenue-positive to the Trea- Hayden Planetarium, to comment that “the highest sury.” To which Maguire replied, “Try explaining form of compliment is [that] people see NASA’s that to Congress.” achievements and they think it’s 20 percent of the Several speakers ticked off space’s many uses: federal budget. They complain about how much crop management, traffic management, disaster we’re spending on NASA to generate what they management, land-use planning, environmental see, and they have no idea how little it is.” For monitoring, antiterrorism. A. P. J. Abdul Kalam, example, “the cost per year of Cassini, all of that India’s 11th president and the aeronautical engineer three billion dollars amortized over the life of the who founded that country’s missile program, boast- mission, is about what Americans spend annually ed that in India, tele-education (2,700 classrooms) on lip balm.” and telemedicine (250 hospitals) are practiced routinely, and that the nation plans to link 100,000 villages by satellite in a “knowledge net.” WHERE WE ARE NOW The European Space Agency’s director general, Jean-Jacques Dordain, proposed having a “Space In some ways, NASA has been a victim of its Day” once a year on which all satellites are shut own success—space is now such a part of our lives off for 24 hours so people can see how much they Below right: The European that we don’t notice it. William Ballhaus Jr., then depend upon them, and suggested, tongue firmly president of the Aerospace Corporation, told “two in cheek, that the World Cup finals might be a Space Agency’s Envisat stories about whether people recognize the impact good day for this. dwarfs its handlers. of space. . . . A CNN reporter asked a soldier in Unfortunately, we will get an involuntary Below left: An Envisat view Iraq, ‘You’re in a net-centric environment, space has demonstration of this as America’s aging fleet of become an integral part of defense, how is space climate-monitoring satellites starts to go off line. of the smoke plumes from affecting you?’ And he said, ‘I don’t need space. As CNN’s chief technology and environment Southern California’s wild- All I need is my rifle and my GPS.’ Another good correspondent Miles O’Brien noted, “By the end fires on October 22, 2007. example is [that] when Dan Goldin was the NASA of the next decade, the U.S. will have 40 percent fewer sensors in orbit because of flat or reduced budgets. Many programs have been canceled; many have been delayed,” which will lead to gaps in the data sets required for climate modeling and prediction. Several speakers bemoaned this state of affairs, and hoped that future administrations would do better. Meanwhile, ESA is stepping up to the task, said Dordain, with Earth missions being its fastest-growing area of endeavor. ESA’s seven-ton Envisat, launched in 2002, is the biggest Earth-observing spacecraft ever built, returning over 280 gigabytes of data per day. A fleet of much smaller craft of some 500 kilograms each will be launched every six months beginning this spring. But NASA’s stumble will be felt—according to Yannick D’Escatha, president of France’s Centre National d’Études Spatiales, “the overall European annual space-science budget is equal to the single Mars-exploration budget [line] in the U.S.” Said Ballhaus, “The fact that weather mod- ESA ESA/A. Van Der Geest

14 ENGINEERING & SCIENCE NO . 1 2 0 0 8 daughters of two of them, now Purdue undergrad- uates themselves, and “one was in our hospital- ity management program, and the other was an English major. Both fine majors, but where’s the engineering spark?” Burt Rutan, founder of Scaled Composites and designer of SpaceShipOne, the world’s first privately built manned spacecraft, pointed out that “all of the billionaires now who are putting money into space—[Virgin’s] Richard Branson, [Microsoft’s] Paul Allen, [PayPal’s Elon] Musk, [Amazon.com’s Jeff] Bezos—were all little kids for Apollo. And

NASA/Alex McClung NASA/Alex that’s where their inspiration came from. Right now, kids are inspired by the next iPhone, not by Climate modeling requires lots of data about any given place at any given time. Hence the exploration or adventure or taking risks. And that’s Afternoon Constellation, better known as the “A-Train” in homage to jazz great Duke Elling- going to hurt us in future generations.” The average age of Apollo’s mission control- ton. These satellites share a polar orbit, with very short time gaps between them. lers was 26, meaning that they were starting high school when Sputnik was launched, Córdova said. “Now at the dawn of a new century that should eling is a utility and push the boundaries of knowledge far beyond ‘the we can’t afford to gap surly bonds of Earth,’ the United States finds fewer it is a national-security students pursuing degrees in science, technology, issue. It’s fundamen- engineering, and math.” Citing various recent tally important that studies, she went on to say that the number of U.S. we plan the future so citizens getting science degrees fell from third in we don’t have those the world three decades ago to 17th in 2004. Forty gaps.” percent of NASA staff are 50 and older; only 4 per- Frank Fernandez, cent are younger than 30. Meanwhile, U.S. jobs DARPA’s direc- requiring a science and technology background are tor emeritus, said growing at 5 percent per year. that with the onset America shouldn’t look for the last couple of of global warming, decades’ influx of foreign students continuing to “We’re going to need, make up the difference, said Tyson, who worried in a decade, the global that our “jingoistic” policies are making it harder There was plenty to see observations to drive the models to make the for them to get into the country. American science between sessions. Here predictions to tell people when to leave. Because has been so fertile because we have always wel- Jean-Jacques Dordain, ESA’s the weather is going to get more active. As more comed the best from abroad, he said. The atomic energy is trapped in the atmosphere, we’re going to bomb and the Apollo program are “American” director general (with red have more chaotic weather.” He added that “some- achievements, but both relied on foreign-born folder); Caltech trustee one needs to make it clear to the policy makers that brains, from Hans Bethe to Wernher von Braun. Jon Kutler (with sun- this is not a ‘nice-to-have,’ it’s a ‘gotta-have.’ We’re Speakers proposed various remedies, including doing this amazing experiment. We’re putting all more industry-academia-government partner- glasses), a member of the of this stuff into the atmosphere, and there’s a lot ships, mostly aimed at college students. Former JPL oversight committee; of controversy over what effect it will have, but Air Force honcho Sega spoke of the new National and Yannick D’Escatha, we do know that it won’t go away by itself for a Defense Education Program, named by analogy to president of France’s couple of hundred to a thousand years. The only Eisenhower’s National Defense Education Act of way we’re going to be able to get the data to tell 1958, and modeled after it. The Act was designed Centre National d’Études us what’s going on and to refine the models to do to bring American students up to speed in sci- Spatiales (CNES), check out the predictions is going to be global sensing. If we ence and math—as Northrop Grumman’s Sugar a replica of one of JPL’s don’t start now, in 10 or 20 years when somebody observed, “It was not just that beeping metallic current Mars rovers. JPL’s finally decides ‘we have to solve this problem,’ we sphere that induced so much hysteria in the free won’t have the capacity to do it.” world, it was the Russian rocket technology that first rover, Sojourner, also This capacity is not just in space, but on the put it there,” and America needed to catch up. full-size, can be seen in the ground, in the form of a new generation of scien- Córdova touted Purdue’s Discovery Park—a 40- background, behind JPL tists and engineers. France Córdova (PhD ’79), acre complex of labs for biotech, nanotech, and the president of Purdue University and an astrophysi- like—as a model for a steady source of funding for director Charles Elachi (red cist with an instrument on ESA’s XMM-Newton high-risk, high-payoff research, which is what the tie) and Sylvie Callari, head X-ray telescope, noted that Purdue’s alumni Apollo program was. Caltech’s own Ares Rosakis of international affairs at include Gus Grissom, Neil Armstrong, Gene described the Institute’s efforts in this field—see the CNES. Cernan, Roger Chaffee, and 18 other astronauts. article on page 34. She recently had the opportunity to talk to the Meanwhile, space is getting crowded. There are

2 0 0 8 ENGINEERING & SCIENCE NO . 1 15 240 commercial communications satellites alone in Boeing’s Whelan proposed a tiered pricing geosynchronous orbit, said India’s Kalam. Speakers system instead of the current one-cost-fits-all from various space agencies reeled off head-spin- approach. Manned missions need the highest ning slates of missions to worlds beyond. The reliability factor—three nines, or 0.999 out of moon is a popular destination—Japan’s Kaguya 1.0—and so one might be willing to pay $20,000 and China’s Chang’e-1 are in lunar orbit right now, per pound. But for satellites, 0.99 reliability would mapping away. India’s Chandrayaan-1 will be suffice, and the cost should drop to $3,000 to launched to the moon later this year. Chinese and reflect that. And for low-value launches of items Indian rovers will follow, then humans. Lockheed like water, fuel, and toilet paper needed for life in Martin’s Maguire quoted Michael Griffin, NASA’s low Earth orbit or on the moon, one could opt for administrator, as saying, “I personally believe 0.9 reliability and pay $200. “It’s a commodity—I China will be back on the moon before we are,” don’t care if you lose it. And in fact, on a strictly and added that the Russians plan a manned mis- economic basis, you only need 0.5 reliability.” sion to the moon in 2025, and a permanent base Rutan opined that “lift capacity started out there somewhere around 2030, despite having one- being so expensive because there was so little of it. sixteenth NASA’s budget. That’s true of every new technology—you do it in onesies, twosies. But now, 25, 50 years down the road, it hasn’t changed much. I think the problem “There’s been nothing much new in propulsion technologies in the last 50 is the folks . . . don’t care that it costs $100 million to buy a booster, because their satellite costs years. But there must be some new discovery still around the corner. You $500 million.” The only solution is to drive down young people at Caltech have a reputation for breakthroughs in propulsion manufacturing costs by making it up in volume, he said, and the only way to generate sufficient technology—let that be your homework assignment for the next week.” demand is “if we go through a time period where

Getting any big project off the ground, and especially out of Earth’s gravitational well, requires international effort these days. As ESA’s Dordain said, “It is always easier not to cooperate, but it is always more difficult to succeed alone.” But cooperation won’t happen unless every party’s national interest is served. Even something as seemingly universally beneficial as free exchange of climate data has pitfalls, said Charles Kennel, former director of NASA’s Mission to Planet Earth, the world’s largest Earth-science program. “Different nations have different conceptions about the value of their environmental intellectual property.” For example, “countries in the Middle East would regard information on the water flows for the next season to be Carreau ESA/C. a matter of national security.” The biggest obstacle to the high frontier is the price of liftoff. It costs $20,000 per kilogram to put a payload in orbit, says Kalam; he’d like to see that figure go down to $2,000 and eventually $200. The problem is the oxidizer, which allows fuel to burn once the air gets thin. Liquefied oxygen takes up nearly 70 percent of the launch weight of a liquid-fueled rocket. Kalam said studies in India show that a reusable single-stage- to-orbit vehicle of some 25–30 tons launch weight is feasible. Such a craft would, once up to speed, scoop up air, liquefy it, and extract the oxygen on Pedoussaut ESA/M. board, accumulating enough to reach orbit. He foresees scaling these designs up to 270-ton vehicles In space, international cooperation is the norm. The with a 15 percent payload fraction. “If you had 20 BepiColombo mission to Mercury (top) consists of two percent payload fraction—10 times current—and spacecraft, ESA’s Mercury Planetary Orbiter and the Japa- reused it 100 times, you’d get about $200 per nese space agencies’ Mercury Magnetospheric Orbiter. kilogram launch cost.” He said that India is working on such systems, and called for international And ESA’s ExoMars rover will include Russian and American cooperation to develop the technologies needed. collaborators. Both missions have 2013 launch dates.

16 ENGINEERING & SCIENCE NO . 1 2 0 0 8 A TRIBUTE TO SI RAMO

A special session of the conference honored Simon Ramo (PhD ’36), a founding giant of the aerospace industry. Session chair Alexis Liva- nos (BS ’70, MS ’73, PhD ’75), the president of Northrop Grumman Space Technology, A computational fluid called him an “entrepreneur, scientist, author, dynamic simulation of musician, philanthropist.” At Caltech, Ramo got dual doctorates in physics and electrical NASA’s X-43 scramjet flying engineering in just three years. “He had more at Mach 7—seven times requirements than he had time,” which once led the speed of sound. On him to take a course and its prerequisite concur- its final flight in 2004, rently. (He passed both.) Livanos described how Ramo and classmate this air-breathing craft Dean Wooldridge (PhD ’36) founded the achieved a record-breaking Ramo-Wooldridge Corporation, “the most Mach 9.6—nearly one- successful electronics corporation in history.” third the velocity needed The duo made the cover of Time on April 29, the focus is on flying the consumer. [Humans] 1957; by then Ramo was Chief Scientist of to reach orbit. don’t cost anything—in fact, they pay to fly. And the Air Force’s ballistic missile program, and these payloads can be easily reproduced in huge Ramo-Wooldridge was overseeing the produc- number with unskilled labor, with tools you tion of the Atlas and Titan ICBMs. (The firm’s already have around the house.” Pioneer 1, launched October 11, 1958, by the “You can now fly nonstop to London for $500 newly formed NASA, became the first privately round trip,” Elon Musk, founder of SpaceX, which built object in space.) In Time’s description, is developing the Falcon reusable rocket, pointed Ramo “‘lets his thoughts bounce around like out. But “the only reason that round-trip flight an errant light beam,’” Livanos quoted. “That from L.A. to London is $500 is because that 747 brilliant light beam has shone brightly at TRW, can be reused thousands of times.” Northrop Grumman [both Ramo-Wooldridge’s Maguire hoped for an outside-the-box solution. successors], and nationally for more than five “There’s been nothing much new in propulsion decades. Si has counseled industry leaders. . . . technologies in the last 50 years. But there must He was awarded the National Medal of Science be some new discovery still around the corner. by Jimmy Carter, and he was awarded the Presi- You young people at Caltech have a reputation for dential Medal of Freedom by Ronald Reagan.” breakthroughs in propulsion technology—let that Livanos and several other conference partici- be your homework assignment for the next week.” pants spoke of how Ramo had mentored them personally—often through one-on-one lunch conversations over simple cheese sandwiches. WHERE ARE WE GOING? Ramo, who was in the front row with his wife, Virginia, did not take the mike, but reflect- Developing cheap orbital access will cost quite ed on his career in a recorded interview with for- a lot of money. The feeling among the speak- mer JPL director Ed Stone. After talking about ers seemed to be that the cash should come from Sputnik, systems engineering, the proper role the private sector as a corollary to the continued of humans in space, and the future of society, commercialization of space. The always-outspoken the conversation turned personal. When asked Rutan led the charge. “NASA should only fund what career advice he would give the next gener- research, not development. When you’re spend- ation, Ramo replied, “If something will interest ing hundreds of millions of dollars to put people you and be right for you, my advice is do what into space using pieces of the shuttle, and pieces comes naturally. . . . When I got to Caltech, of Apollo, you’re dumbing down a whole genera- I found it was a science school. It didn’t really tion of new, young engineers who are told, ‘No, teach engineering, it taught the science underly- you can’t take any risks, you’ve got to do it the way ing engineering . . . at a period of time where we know will work.’ . . . Having the government engineering was about to be changed greatly, repeat Apollo 50 years later is just silly.” Musk felt because of the scientific breakthroughs taking that the potential was indeed there. “If we have a place.” He got a job at General Electric upon base somewhere, on the moon or Mars, hopefully graduation, and “the older engineers . . . did not a tiny growing civilization, the transport back and know the new physics. So I had opportunities forth will be a multitrillion-dollar industry.” right off the bat to do some things. . . . It was David Thompson (MS ’78), chairman and CEO an acccident. I was just lucky.” —DS of Orbital Sciences Corporation, which builds and

2 0 0 8 ENGINEERING & SCIENCE NO . 1 17 launches satellites, charted the vital signs of a large and thriving space industry—$82 billion in commercial satellites and ancillary services worldwide in 2007, versus $42 billion in defense and $27 billion in all other areas, including manned spaceflight. Satellite TV revenues are now double the total U.S. national security space budget—$52 billion versus about $25 billion—“not bad for a business that didn’t even exist 15 years ago.” GPS equipment revenue, “the success story of the next decade,” is bigger than our missile-defense budget ($10 billion versus $9 billion). And commercial remote sensing (think Google Earth, through which, as Rutan remarked, “in the last few years, with free down- loadable software, you’ve been able to do what only a few analysts at the CIA used to do.”) is approach- ing a billion dollars a year. “While Apollo and its predecessors drove the These three views of Hurricane Katrina give an inkling technology development that launched the first of the wealth of data that needs to be integrated into generation of space businesses, now technology climate-prediction models. From left: NASA’s Aqua satellite development often flows in the other direction,” Thompson concluded. “It is time for the private recorded sea surface temperatures on a three-day average. space sector to repay the debt to NASA that got it The Tropical Rainfall Measuring Mission, a joint project started in the 1960s.” of NASA and the Japan Aerospace Exploration Agency, Several speakers speculated about the Next Big tracked rain-cell intensity in three dimensions. And NASA’s Thing. Would it be tourist hotels in low Earth orbit? Mining helium-3 on the moon and export- QuikSCAT satellite depicted wind speeds in color and wind ing it to Earth to produce clean energy by nuclear directions with small barbs. fusion? Orbiting arrays of solar panels that would beam their harvested power down? But all agreed that an airtight business plan would be required to structures like space stations, or spacecraft destined start such a venture, and none were in the offing. for Mars and beyond. NASA is already planning Boeing’s Whelan sketched out an infrastructure to use a similar approach for moon base construc- for space, using DARPA’s Orbital Express as the tion, launching the lunar lander with just crew and exemplar. A collaboration with NASA, the Air cargo and then fueling it aloft, allowing it to ferry Force, Boeing, and other firms, Orbital Express a lot more weight per trip. Said Whelan, “I may consisted of two vehicles, Boeing’s Autonomous not be able to reduce gravity, but I can improve the Had there been a third Space Transport Robotic Operations, or ASTRO productivity of every pound I send into space.” (the servicer, named in homage to the family dog Not surprisingly, the future of space science was spacecraft watching, in the 1960s prime-time animated TV series, The a hot topic. Our own planet will continue to get Orbital Express’s ASTRO Jetsons) and Ball Aerospace’s NextSat (the cli- close scrutiny, although Berrien Moore III, then (two solar panels) and ent). For four months in 2007, the two spacecraft director of the University of New Hampshire’s NextSat (one solar panel) demonstrated that they could perform such feats as Institute for the Study of Earth, Oceans, and rendezvousing with each other, docking themselves, Space, remarked that global warming in and of might have looked like this and executing service missions such as refueling itself is “not what the body politic is interested while docked. and component swap-outs—all using only their in. After all, if we double the amount of CO2 in onboard computers the atmosphere, it will lead to a three-degree, on and navigation sys- average, temperature rise. Many people move from tems, with no human New Hampshire to Arizona for that three-degree intervention. Whelan global warming. What people are interested in proposed that fleets of are changes in severe events or extreme events.” unmanned transports With advances in computing, “we are learning to carrying fuel and spare put the pieces together—we can link wind fields parts would one day with rainfall fields and sea-surface temperatures to rendezvous with auto- understand the mechanisms of hurricanes.” He

mated service vehicles. pointed out that even though atmospheric CO2 These sophisticated levels do not cause Earth’s glacial cycle—that’s an robots would do orbital thing—they have closely tracked the tem- whatever maintenance perature records for the last 400,000 years. Earth was needed on clients is now in an interglacial period, which means that

near their orbits, and CO2 levels should be about 350 parts per million even assemble large by volume. Instead, “we’re now at 450, 100 over

18 ENGINEERING & SCIENCE NO . 1 2 0 0 8 the recorded high, and if we convert to a clean energy system the IPCC [Intergovernmental Panel on Climate Change] thinks we can stabilize it at 550. Business as usual puts us over 1,000. Even

if CO2 were not a greenhouse gas, this would be a major disruption in the carbon cycle, and it would be worth studying very carefully.” Several missions, including JPL’s Orbiting Carbon Observatory and Japan’s GOSAT (Greenhouse Gases Observing Sat- ellite) will track carbon’s sources and sinks, and the decadal plan is to “orchestrate that with other missions that look at changes in vegetation, changes Moore quoted what poet Archibald MacLeish wrote in The in radiation balance, changes in aerosols, changes New York Times on Christmas Day, 1968, upon seeing in ozone to begin to get a handle on changes in the first photo of our home planet ever taken from deep climate.” The radiation balance—how much heat we get from the sun versus how much we lose back space: “To see the earth as it truly is, small and blue and beautiful in that eternal silence where it floats, is to see Saharan dust storms to space—is critical, Moore said, because climate change is altering the rate at which the planet soaks ourselves as riders on the earth together, brothers on that like this one have been up heat. Bright, reflective ice and snow is becom- bright loveliness in the eternal cold.” The picture had revealed to deliver trace ing dark, absorbent open water and exposed soil, been taken on December 22 by Apollo 8 astronauts Frank nutrients, including iron, for example. “We see that in the Arctic sea ice. Envisat in 2007 showed us that the long-searched- Borman, James Lovell, and William Anders as they rounded to vast areas of the ocean for Northwest Passage has finally appeared.” the far side of the moon. that would otherwise be As for other worlds, Jonathan Lunine (MS ’83, much less fertile. Through PhD ’85), professor of planetary science and phys- global satellite monitor- ics at the University of Arizona and an interdisciplinary scientist on NASA/ESA’s Cassini/Huygens ing, “we’re beginning to mission, talked about its exploration of Saturn’s understand couplings that moon Titan as a paradigm, both as a model of we didn’t expect,” said international cooperation and an intrinsically Moore. This shot of north- interesting destination. He called Titan a “Once and Future Earth” because its atmospheric chem- west Africa and the Canary istry—hydrocarbons and nitrogen, but no oxy- Islands was caught on July gen—resembles prebiotic Earth’s, and at the same 24, 2003, by NASA’s Terra time Titan may presage Earth’s eventual fate long after humans have died off or moved on. “As the satellite. sun’s luminosity grows [Earth’s] temperatures will rise, the oceans will evaporate into the stratosphere, and the water will be photolyzed, just as methane is on Titan.” He described a JPL proposal to explore A balloon-borne survey of Titan could send back data for Titan “in a leisurely way, with imaging that covers months instead of hours. hundreds of thousands of square kilometers, rather

2 0 0 8 ENGINEERING & SCIENCE NO . 1 19 Backdropped by Earth’s horizon and the blackness of space, the International Space Station is seen from space shuttle Atlantis as the latter departed on February 18, 2008.

than just 150,” riding the gentle winds in a hot-air percent of the federal budget, or “about 15 cents balloon lifted by the waste heat from the radioiso- per American per day” for the foreseeable future. tope power source that would provide the electric- “In Washington, it’s very difficult to get programs ity for the instruments in the gondola. started, but it’s even more difficult to get them Looking beyond our solar system, JPL’s Michael stopped.” Given that, he said, “the president’s goal, Werner of the infrared Spitzer Space Telescope return to the moon by 2020, is a rational goal in noted that the visible-light follow-on to NASA’s constant dollars.” The Hayden’s Tyson agreed that Great Observatories program, the James Webb more could be done with the manned program’s Space Telescope, is slated for launch in 2013 and share of the money than what we have to show will have 50 times the light-collecting area of the for it recently. “We designed Mercury, Gemini, Hubble. With this and other missions, we will Apollo, and Skylab on that budget. We got out of one day answer the Big Questions: Where do we low Earth orbit on that budget. And ever since, come from? Are we alone? Some of the answers we’ve been driving around the block on the Shuttle, are already coming in, he said. “How prevalent are boldly going where hundreds have gone before.” Over 1,000 people in more planetary systems around other stars? We can say Griffin said that even Mars is reachable on that than 17 countries are with confidence that they are either prevalent or budget. Assuming that the manned portion ($9 very, very prevalent.” billion of NASA’s $14 billion per year) continues to developing the James Webb The running controversy in the academic go to the International Space Station until around Space Telescope. Some community over the merit of manned versus 2020, and then gets rolled over into developing a team members pose with a unmanned missions was explored. NASA admin- lunar base—which would cost a few billion dol- full-scale model at NASA’s istrator Michael Griffin debunked the widely lars a year to sustain—there would be five or six held notion that NASA’s budget peaked during billion a year left over to begin working on Mars. Goddard Space Flight Cen- the Apollo years, and has been in steady decline At that rate, we could get to Mars in about 15 ter in Greenbelt, Maryland. ever since. Although it is true that NASA spent years, because most of the heavy lifting—literally, more money in the the development of the Ares V booster—will have few years when it was been borne by the moon program. “We could be buying large pieces of launching in 2035 and landing in 2037. And so by expensive hardware the time we are celebrating the 100th anniversary like Saturn Vs, he said, of Sputnik, we can be celebrating the 20th anniver- if the budget is aver- sary of landing on Mars.” aged over the length In a rousing closing keynote speech, Tyson stated of time it takes to see a that manned space flight was NASA’s top priority. big project like Apollo “We in the academic community have this delu- or Cassini through to sion that NASA is our own private science fund- its end—say a 10-year ing agency. But it has never been that.” In fact, running average—the he said, NASA’s science budget began at around flow of money has 10 percent in the ’60s and has grown to about been remarkably 30 percent today. “But most of the money was steady. never directed at science.” He pointed out that the Griffin sees NASA’s manned program is spread out across the country, funding as remaining where the representatives who vote NASA’s budget constant at about 0.6 live, and that the citizens who elect those repre-

20 ENGINEERING & SCIENCE NO . 1 2 0 0 8 sentatives are interested in people. “I don’t think anybody is ever going to name a high school after a robot. Yes, people love the rovers. I’m told the number of hits to the JPL website monitoring the rovers in the two weeks after they landed exceeded the total world traffic in Internet pornography.” But, he said, “I would bet you that if humans were on their way to Mars at the same time, the rover site would go unremarked-upon.” In the early days of manned flights, “every mission was an advancement on the previous one, which gave the media something to talk about. . . . That is a truth that’s never recognized by the naysayers of the manned program—science is piggybacking on the manned program, and always has been. And I have not Spacecraft old and new: Up front is a mockup of the Phoe- been given reason to presume that that will ever nix Lander, slated to touch down in the north polar region change. So if you have a healthy manned program, of Mars on May 25. In the background is Explorer 1 atop science will be riding on its back. Maybe not as much as we academics would like, but it’s there.” its second-stage Jupiter-C rocket. While agreeing that sending mankind to Mars could be done, Tyson wondered if it would be done. highway system. (It, like Apollo, was a product He summed up our species’ other great cultural of the Cold War—highway construction began in investments of human and financial capital, and earnest in 1956 to move troops and tanks around stated that there were only three drivers powerful the nation in the event of a Soviet invasion.) “The enough to motivate them. “War. That is the big- Promise of Economic Return.” Here he listed gest driver there ever was. There is always money the voyages of Columbus and Magellan, and the for war.” In this category he put the Great Wall of Tennessee Valley Authority. “Praise of Power.” China, the Manhattan project, and the interstate This could be civil or religious, he said: pyramids, cathedrals, or castles. “If you’re a wealthy nation, you can do a billion-dollar mission. But can you Clockwise from the top: 1. Conference co-organizer do a 10-billion-dollar mission? I don’t know. Dwight Streit, vice president, Northrop Grumman Space There’s a threshold. Did we build the Supercon- Technology. 2. Conference cochair Alexis Livanos, presi- ducting Supercollider? No. Because there wasn’t a weapon at the end of the supercollider. There dent, Northrop Grumman Space Technology. 3. (from weren’t diamond mines at the end of the supercol- right to left) Conference co-organizer Ares Rosakis, director lider. We didn’t see the face of Jesus at the end of the Graduate Aero- of the supercollider.” To sustain a project with a nautical Laboratories at Mars-sized price tag, he argued, only one of the Big Three will do. “All it would take would be one the California Institute of message from Beijing: ‘We’re going to put military Technology; A. P. J. Abdul bases on Mars.’ Badda-bing; we’d be on Mars in Kalam, the 11th presi- nine months.” dent of India; and Kent Tyson ended the conference with slides of various technologies at infancy and maturity: Karl Kresa, chairman emeritus Benz’s first motorcar and a high-end Mercedes, a of Northrop Grumman Wright Flyer and an Airbus A380, and so on. “The and chairman of Caltech’s first one looks quaint and should go in a museum. . . . But the Saturn V, 36 stories tall, does not look board of trustees. 4. quaint. You have the urge to genuflect when you Ronald Sugar, chairman of walk by it and ask, ‘How the hell did we achieve Northrop Grumman. this thing?’ . . . The challenge today is to work at something that will make the Saturn V look quaint. That’s what’s going to have to happen to enable the next generation of space exploration.” Whether the means to get us into orbit will under- go the same rapid technological evolution as the telephone, the mouse, the PC, and other workaday devices, ushering in a bright new era in space, only the next 50 years will tell.

PICTURE CREDITS: 13, 17, 19, 20— NASA; 12 — NASA/JPL-Caltech/U. of Arizona; 13 — USGS, NASA; 18 — Boeing, NASA/SVS; 19 — NASA/JAXA, NASA/JPL; 20 — NASA/GSFC; 15, 21— Bob Paz

2 0 0 8 ENGINEERING & SCIENCE NO . 1 21 Auctioning off the FCC’s Crown Jewels

By Elisabeth Nadin

means fewer towers and The 700 MHz spectrum therefore less expense consists of several blocks to the winning bidders. dispersed across aban- To cell-phone users, it doned analog television may also mean good- bye to roaming charges. channels. Hierarchical Auction 73, as the 700 Package Bidding (HPB) MHz auction is offi- was tested on the C block, cially known, was the which offered a total of first chance for newbies like Google and Cox 22 MHz in 12 Regional Cable, who each fronted Economic Area Groupings the money the FCC (REAGs) that together required to enter the auction, to break into would span the nation. the wireless scene. They (Blocks in gray were auc- would bid alongside tioned before Auction 73; entrenched giants like the D block was set aside Verizon and AT&T. Caltech professor of for public safety announce- economics Jacob Goeree ments but failed to meet monitored the auction the FCC minimum reserve closely, focusing particu- price.) larly on one segment of the spectrum called the C block. The auction It’s been called beachfront property . . . the last action was centered here because the C block—two big slice of the spectrum pie . . . the crown jew- bands totaling 22 MHz of the spectrum—would els of the Federal Communications Commission be sold under a new scheme: companies could bid (FCC). It is the 700 megahertz (MHz) frequency either for any of 12 large-region licenses or for a band, which actually runs from 698–806 MHz package deal to win coast-to-coast coverage. By and until now has been the exclusive domain of the end of the auction, which dragged on for 261 broadcast television. On January 24, as a step on rounds spanning 38 weekdays as bidders haggled the road toward ubiquitous digital television, the over tiny portions of the 700 MHz offerings, the FCC began auctioning off licenses for other uses C block had generated about a quarter of the of swaths of this band in what was called the most $19.6 billion total that the FCC reaped in Auction significant airwave auction in U.S. history. 73—the most money the federal treasury has ever The 700 MHz band is a hot commodity, espe- earned in a single auction. All the bids had been cially for wireless companies, because the signal placed anonymously, but before the FCC finally penetrates walls. Each tower broadcasting in this announced the winners on March 20, those who range can cover at least four times as many square had followed the daily action speculated that Verizon miles as conventional cell-phone towers, which and AT&T were locked in a bidding war with

22 ENGINEERING & SCIENCE NO . 1 2 0 0 8 Google for nationwide CRUNCHING NUMBERS coverage, driving prices up. All that had been During the two years preceding that FCC auc- known for sure was tion, you would ﬁnd students sitting at computers

Image: CoverageRight. Image: that someone was cast- in the Social Science Experimental Laboratory ing bids that topped $2 (SSEL), in Baxter Hall’s basement, weighing their billion for the nation- options, placing their bids, and waiting for the next wide package as early round. SSEL director Goeree would stride around as the fifth round. the room watching the students in action, or bend Goeree was person- his tall Dutch frame over his own computer screen, ally invested in the tracking each bidder’s move. He had outlined the action around the stakes—how much money each student stood to Above: Current coverage at the 850 MHz and 1900 MHz C block, as he had make if their bid won—and exhorted them to keep frequencies—the primary mobile communications bands in designed the bid- their profit margins as high as possible. Goeree ding mechanism that wanted to know, could the bidder who stood to North America—leaves much of the country with no signal. was used. Called profit most actually win the auction? Time and Below: The 700 MHz signal propagates much further, Hierarchical Package again the answer was yes. necessitating fewer cell-phone towers. Bidding (HPB), it’s Goeree had no problem attracting students to a relatively straight- run his experiments. “I heard about it from other forward system that students,” says physics major Justin Chen, who lumps individual pieces into larger units that might earned enough in the FCC experiments to buy prove more attractive to someone who his longboard, on which he’s often spotted roll- would rather buy a whole pie instead ing through the campus. (Chen is no stranger to of a few slices. In Auction 73, it meant strategic decision-making: he and a friend at Har- that either some companies—Verizon vard concocted an equation to help them decide or AT&T—could win one or more of whether it’s better to just wait for a bus or to start the 12 regional licenses and fill in walking. Their answer, which made worldwide their coverage gaps, or another news, showed it’s almost always better to wait.) company—Google—could make “It’s generally straightforward,” remarks Chen. a big splash by winning the 12- “The experiments are designed assuming that all license package to cover the whole the players want to make as much money as they nation. “The HPB auction rules let the can.” Before each round, each student bidder read market discover the best allocation, and from their screen what licenses they were interested how things get packaged, which takes in and how much winning them would be worth: the heat off public officials who would Their earnings from the auction were proportional have to respond to lobbying pressures to the difference between those values and their on these issues,” Goeree says. HPB bids, if their bids won. The better the bidder, the was deemed by the FCC to offer more he won, and in this manner Chen scored his the best chance for new greatest win of $140. A friend once made $303. entrants or new business Not bad for two hours. models to penetrate Even though the students didn’t actually know the wireless grid. what they were supposedly buying, they com-

2 0 0 8 ENGINEERING & SCIENCE NO . 1 23 peted with each other so realistically, says Goeree, licenses into the hands of those who value them that their bidding mimicked professional auc- most. And the FCC came out ahead, too—it no tion behavior. Their motivation, after all, was the longer had to figure out which applicants qualified; same. “You start thinking about your strategies in each could bid as high as they decided a license the game—how you’re going to bid to make more was worth to them, and the FCC would reap the money,” Chen says. He discovered that the main profits. These auctions employed simultaneous problem with auctions is that “you can’t really multiple bidding for individual licenses—all the lose money, but you can pay too much for some- companies interested in a license would bid on thing.” Auctions are supposed to run such that every license they were interested in, and then stay the winning bid goes to the person who values the in the bidding as long as they had the money. prize most. But in an ordinary auction format, the The first FCC auction back in 1994 was consid- FCC would have no way of making sure that hap- ered a great success—the New York Times termed it pened, even though they are charged with award- the “Greatest Auction Ever”—and even then Caltech ing licenses in an efficient manner that serves the researchers played a pivotal role. Caltech’s Johnson public interest. Professor of Business Economics and Management, Preston McAfee, helped design the auction together with Stanford University professors Paul Milgrom AUCTIONING THE AIRWAVES and Robert Wilson. Another Caltech scientist, Charles Plott, the Harkness Professor of Economics Back in 1927, the Federal Radio Commission and Political Science, had been testing the FCC’s faced a straightforward task in granting airwave system and discovered that the bidding software was access—the demand was low enough that every flawed shortly before the auction was supposed to applicant got a license. By the time the FCC start. Plott impressed the FCC, and the industry, by took over seven years later, television was on its providing the FCC with a manual backup system way. Several parties competed for each license, he created with Antonio Rangel (BS ’93), then an and the FCC had to weigh which one had the undergraduate student and now an economics pro- public’s interest best in mind. The winner often fessor at Caltech. sold its license for a profit, and losers appealed and Another snafu the FCC faced in the auctions won, and the FCC found itself mired in lawsuits. was bidding collusion. Companies weren’t allowed It took a lot of time and tax dollars to grant a to communicate directly but they could signal license. License fees were fixed at a low rate, and through the bidding process how to divide the the FCC sure wasn’t pulling in any money for all market. With bids that reached over $100 mil- its work. lion, bidders could use the many zeros in the bid amounts to signal information that might help in an attempt to keep prices down. The FCC tried to “The experiments are designed assuming that all the players want to make as fix this by imposing predetermined bid increments, but the suspicion of bidder collusion continued to much money as they can.”—student volunteer Justin Chen overshadow later auctions. The problem was finally resolved in Auction 73, when the FCC decided to use an anonymous bidding procedure. Until the close of the auction, no one but the bidder knew Imagine what happened as the telecommuni- who had placed the highest bid in a round. cations industry grew. The FCC now regulates The FCC’s simultaneous bidding mechanism, interstate and international communications by used for over a decade and copied around the radio, television, wire, satellite, and cable. In 1982, world, is generally considered a great success. But the agency switched to a semiregulated but arbi- even prior to that first auction in 1994, some trary license-granting system: a lottery, allowing the economists saw room for improvement. They wor- FCC to award one applicant and avoid litigation ried that companies interested in winning certain from the losers. But then came a flood that would combinations of licenses, like a package that might have made Noah blanch: applicants filed under serve the entire East Coast or establish nationwide multiple names to increase their chances, and coverage, might be hurt in the license-by-license opportunists realized they stood to make a killing if competition the FCC was organizing. Several they happened to win—they could sell the license Caltech professors, including John Ledyard, the at a vastly inflated price to someone else who really Davis Professor of Economics and Social Sciences, wanted it. This was especially true for cellular pioneered a different approach, one in which bid- communications, and wireless lottery schemes were ders could place bids on individual licenses as well called “the number one investor fraud in the coun- as combinations of licenses. Their research had try.” The FCC seemed like the biggest loser under convincingly shown that efficiencies and revenues the lottery system. of the FCC auctions were reduced because bidders A little over a decade later, the FCC switched to hesitated to incorporate synergistic values into their a competitive bidding system. The system makes bids for fear they would end up being in a bad spot sense to economists, says Goeree, because it gets financially when competing fiercely for a desired

24 ENGINEERING & SCIENCE NO . 1 2 0 0 8 wins for the bidders. Most of all, it alien- ated bidders with its intricacies. Goeree and Holt tested related auction designs, but these didn’t fly well either and were still too complex. Goeree recalls thinking that after having discussed In this HPB scheme involving pies, bidders competed on three levels. The winning bids, in red, maximized profits by the possibility of combining winners in levels two and three. The level-one bidder, who stood to take home 20 pies, couldn’t outcompete combinatorial auc- bidders on level two, who wanted packages of five pies. Likewise, the level-two bidder for the second “block” of five pies tions for over a decade, the FCC might opt lost to those on level three, who bid on the individual pies of that package. out altogether. So he decided to create a new method. “We had a very simple idea package but winning only part of it. Their findings for how to do it,” he says. “First, imagine you formed the impetus for the FCC to build a new construct a hierarchy of packages by dividing a system that incorporated combinatorial bids. large nationwide package into two pieces, dividing each of those into two pieces, and on and on, all the way down to the smallest geographic regions BUILDING A BETTER AIRPLANE that could not be divided further.” Unlike previous combinatorial auctions that were considered When the FCC approached Goeree and his col- NP-hard, such a hierarchy makes it trivial to find league Charles Holt at the University of Virginia in the best allocation: simply compare the revenues 2004, it asked them to test the combinatorial auc- that result from selling in one hierarchy level to tion the FCC had already built. Goeree recalls, “It the next, starting at the bottom and recursively was as if they were saying, ‘Before we bring people solving to the top. The challenge Goeree and Holt on board, please fly our plane around.’” Like test- faced was determining appropriate prices given that ing a scale model of an airplane in a wind tunnel, package bids would be placed on many different experimental auctions allow economists to control levels. After they solved how to “trickle down” the all the variables. “We know everything because we excess amount of a winning package bid at a higher induce it ourselves,” Goeree remarks. level by imposing “taxes” on lower-level licenses, The FCC wanted a method in which all buyers Goeree decided to call the FCC and present HPB could compete equally and the commission would as a viable alternative. make the most money in the process. It turned out In his initial tests of the HPB auction, Goeree the combinatorial auction the FCC had devised grouped the available licenses into packages in a for the 700 MHz auction was too complex. “In three-tiered hierarchy, as shown at the top of this computer science it’s known as the knapsack prob- page. Say the HPB is auctioning off pies. On the lem,” says Goeree. “When you have a knapsack of top level, level one, you stand to win all 20 assorted finite volume and you can choose among objects pies in one fell swoop. On level two, there are four of different sizes and values, how do you pack it to packages, each consisting of five different flavors. Caltech students compete maximize its value?” Or, in mathematical parlance, On level three, you can bid on the 20 pies sepa- in earnest in HPB trials. the problem was “NP-hard”—it exploded expo- rately. Now say you’re throwing a party for 200 nentially because there people—well, you might as well go for the gusto were far too many on level one. But if you merely want one pie to potential combina- take home for dessert, your choice is equally clear: tions on which to bid, bid on level three. There could be 19 more people in this case too many like you who also want only one pie, and all your bandwidth licenses bids together might just win out over the level-one spread over too many bidder. The intermediate level, level two, may geographic regions. appeal if you own a small diner and want to serve a “We tested the plane few different options. And if you wanted a package the FCC built and it of five from level two and one more apple pie from didn’t fly that well,” level three, well, you could even bid on both levels. says Goeree. It didn’t As the bids roll in, the party-throwing bidder on maximize profits for level one might find he can’t afford to outbid the the FCC or potential personal pie eaters on level three. He might revise

2 0 0 8 ENGINEERING & SCIENCE NO . 1 25 his strategy and start bidding on three packages million to make up the $0.4 billion difference, and on level two and two more pies on level one. The the bidders on level one would be told they could HPB format not only allows bidders to decide what sit tight until the next round. There is no need to sort of bid suits their needs, it provides flexibility as do your own calculations; the bidders just check the auction progresses. if they can afford the suggested bid. “It solves the In the view of student volunteer Chen, “HPB is complexity for them,” says Goeree. It also meant better because you can win two licenses combined that if each bidder at level two followed the advice as a package for less than what you might bid on in unison, they would all move on to the next the two individually.” So if you want an apple pie round. Of course, whoever couldn’t fork over the and you’re clearly winning it, and you also want dough would get shut out. a pecan pie for which there’s a lot of competition Although the opportunity to win it all in one that you can’t afford to outbid, you could lump fell swoop is a strong appeal of Goeree’s packages, your resources together for a stronger bid on a the FCC saw HPB as an opportunity for smaller package that includes both. players to merge forces at a lower level and overtake one giant bidder at the top. In an October 2007 public notice, the FCC declared, “The HPB auction format was chosen in part because it mitigates issues inherent in some other package bidding for- mats that give bidders interested in large packages an advantage over bidders interested in individual licenses.” The FCC also liked Goeree’s calculation tool and the way the software seemed to prohibit collusion. “In fact, we will use HPB in part because the mechanism for calculating [prices] is significantly simpler than other package bidding pricing mechanisms,” it stated. “In addition, we find that . . . HPB procedures in general strike a careful balance between permitting bidders adequate bidding flexibility and discouraging insincere and anticompeti- tive bidding behavior.”

HPB IN ACTION

REAGs for the C block It became immediately clear that the option to Although the FCC ultimately chose to use a buy the C block in its entirety was extremely valu- split the country into 12 two-tiered system, Goeree says his testing showed able. In July 2007, Google chairman and CEO regions. In the two-level that even this was more efficient than the previous Eric Schmidt had written an open letter to the HPB format that the FCC format. At the bottom level, 12 individual licenses FCC chairman offering a deal: the promise of a corresponds to 12 geographic regions that the minimum $4.6 billion bid on the national package chose, bidders could win FCC designated—Region 1 is the Northeast, for in exchange for open access to the wireless network all 12 regions in one fell example; Region 4 is the Mississippi Valley; Region that would be set up on the C block, regardless swoop or bid on them 12 is the Gulf of Mexico. The top level, level one, of who won it. Open access means any wireless individually. was forged into a three-package deal: a 50-state customer can download any software and use it on grouping of eight of the 12 licenses, another pack- the device of their choice, and the service provider age of two covering Pacific island territories, and a would have to abide. This doesn’t just mean ring two-license Atlantic package combining the U.S. tones, it applies to anything you might use your Virgin Islands, the Gulf of Mexico, and Puerto cell phone for: image-processing software for cam- Rico. era phones, e-mail software for Blackberries, maybe The bidding kicked off on January 24, with one an iTunes knockoff for your Kyocera phone. In round per day. (On day 10, the FCC acceler- the recent past, a closed market meant that AT&T ated the bidding to five rounds per day to speed could connive with Apple for exclusive rights to up results.) At the close of every round, Goeree’s providing iPhone service, and Apple could void the software totaled up the money bid at each level. phone’s guarantee if users hacked their gadget by In a two-level system, let’s say the bids at level two installing non-Apple software. totaled $0.8 billion after day one, and at level one It was also well known that Google was devel- the top bid was $1.2 billion. The software then oping its first cell phone. If Verizon Wireless advised bidders on what their next move should be ended up a big winner in Auction 73, Google’s if they wanted to stay in the game. Thus, if there open-access bid assured that Verizon would have were 12 bidders on level two, they would each be to provide service to anyone who wanted its plan alerted to increase their bid by a little over $33 on a Google phone. Indeed, Google’s demand

26 ENGINEERING & SCIENCE NO . 1 2 0 0 8 prompted Verizon, months before the start of the colleague Holt put it, “The lead switched about the auction, to grudgingly commit to open its network same number of times as in the final quarter of the to wireless devices, software, and applications that Super Bowl.” the company did not offer. This would happen As it turned out, Verizon and AT&T spent even if Verizon did not win the national package the lion’s share—$16 billion—of the FCC’s total —because of the way the auction is structured, earnings. Verizon won C-block coverage of the as long as C-block bids topped $4.6 billion, the continental U.S. and Hawaii, while AT&T swept winner(s) would be obliged to allow their custom- up most of the 12 MHz sold as the B block. It ers to use any mobile phone they wanted and allow seems likely that Google never really wanted to run outside applications to run on their network. a national cellular network, and was happy to call On day six of the auction, electronics bloggers it quits as soon as it secured what it had entered the across the world celebrated. “There was a brief, game to get. And apparently no one else wanted or tense pause in the bidding this morning, but some needed nationwide coverage once open access was anonymous giant telecom company (Google, per- guaranteed. But even though HPB did not bring haps?) has just pushed the price of the 700 MHz C a new entrant to the wireless market, Goeree says, block over the FCC’s reserve price of $4.6B—and “this auction is a winner because of the open access. the rest of us straight into the promised land of It never would have been possible without the 50- open access,” one wrote. “Yep, January 31, 2008, state package.” Round 17 will be the day to remember.” Goeree sees room for improvement, however. The technorati were abuzz, issuing daily progress The auction was implemented with a mix of hier- reports on Auction 73. As it turned out, bidding archical package bidding for the C block and the on the C block at level one ceased shortly after the FCC simultaneous single-license bidding for other open-access benchmark was passed. The bids on blocks, which were offered in a wide dispersion of level two rose to nearly $5 billion as bidders ended license sizes. This dispersion, together with FCC up duking it out over licenses for the 12 individual “activity rules,” made it difficult for firms to reenter regions. the bidding on the C block after they had started The auction as a whole drew to a close on March bidding on other licenses, which may have created 19, and the FCC waited one day to announce the some inefficiencies. winners. The leading bid for the C block had shift- Overall, the auction is evaluated to be a success. ed back and forth only twice between the cumula- The main outcomes of the 700 MHz auction, tive prices for the individual licenses and bids for entry and open access, would have been virtually the national package, demonstrating, according impossible in the license-by-license competitions to Goeree, that the market was determining how the FCC has organized so far. “Competing for the spectrum would be most valued. As Goeree’s open access by pushing prices over a $4.6- billion hurdle is simply too risky when bidding on individual licenses,” Goeree says. “You may end up Goeree, pictured here, paying high prices for a subset of licenses that can’t hopes that the HPB design guarantee a profitable business plan.” will transform wireless The current chairman of the FCC, Kevin Martin, called the auction the FCC’s “transforma- markets in other countries tive auction.” When asked about his legacy at the as well. He recently FCC after the Bush administration leaves, Martin visited Taipei, where he responded, “I certainly think that the success of this auction, the success of raising more money presented the details of than the commission ever raised before in any the mechanism to a group auction, the success of moving forward with a of scholars involved with more open platform that will transform the entire designing Taiwan’s own 700 wireless industry, is going to be a significant accomplishment.” MHz auction.

PICTURE CREDITS: 22, 26 — FCC; 25 — Jacob Goeree; 23, 25 — Doug Cummings

2 0 0 8 ENGINEERING & SCIENCE NO . 1 27 When Cells of a Feather Don’t Flock Together

By Marcus Woo In a scene worthy of a horror movie, pill-shaped remains a mystery. Cells choose different fates, bacteria cells glow green in the darkness. The even when they’re genetically identical and grown ghostly cells squirm, grow, and divide, expanding in the same environment. As creatures become their reach across the screen. Some shorten and more complex, so do the choices. In mammals, for turn into white spores, entering a dormant state example, the power of embryonic stem cells lies in from which they can come back to life decades— their ability to turn into anything, from blood to even centuries—later. Others temporarily glow red muscle to skin. Understanding the diversity of cell when they choose yet a different fate. But not all behavior could lead to new ways to attack cancer of the cells become spores, and only a few percent or develop new drugs, Elowitz says. If you want a ever turn red at a given time. Michael Elowitz, drug to affect a cell a certain way, it helps to know assistant professor of biology and applied physics how it behaves. and Bren Scholar, and colleagues created these cells Elowitz’s lab is peeking under the hood of the to explore the fundamental questions of how and molecular interactions within the cell, where when they decide to change. Applying the innova- networks of genes, proteins, and other molecules tive approaches that won him a 2007 MacArthur work in concert to ensure the cell does what it’s Fellowship, the so-called genius grant, Elowitz supposed to do. Called genetic circuits, these wants to understand how a cell’s genes work to networks “form the foundation of all biology,” says make these decisions. Gürol Süel, a former postdoc in Elowitz’s lab and Even for simple organisms like Bacillus subtilis, now an assistant professor at University of Texas this transformation process, called differentiation, Southwestern Medical Center. “Any biological

28 ENGINEERING & SCIENCE NO . 1 2 0 0 8 The colors of B. subtilis demonstrate variations in cell gene switches on and off. One of the genes was behavior from random fluctuations in the cells. Research- engineered to light up when on, so that, when the ers engineered cells to glow green and red depending on researchers inserted their homemade circuit into a bacterium called Escherichia coli, the cells pulsed. which of two nearly identical genes were expressed. Fluc- Elowitz and Leibler called their circuit the repres- tuations in gene expression, called intrinsic noise, cause the silator, a combination of the words “repressor” and two genes to be expressed differently. Instead of a uniform “oscillator.” yellow from equal parts green and red, the cells glow in The repressilator was an exciting proof-of-prin- ciple experiment, showing that the genetic circuits shades of yellow, green, and orange. underlying cellular behavior can ultimately be understood and manipulated. Researchers had tinkered with genetic circuits before, but this was the first made entirely from scratch that also had a dynamic function. “We were very happy at first— wow, we were able to create this oscillator,” Elowitz recalls. But they soon noticed something odd. To study the repressilator in detail, they had made black and white movies of the E. coli cells. Each film began with a single cell, dark and faintly visible, at the center. The cell began to divide, process you can think of occurs as a result of inter- and its offspring pulsed, casting an eerie glow on actions between biological molecules. There’s no their neighbors. But none of the cells pulsed in gene or protein that acts on its own—everything is sync with the other—even though all of the cells interacting.” had identical genes. “If the oscillator were behav- ing exactly the same in all the cells, they would all remain in perfect synchrony,” Elowitz says. “But A NOISY CLOCK they clearly don’t—they get out of synchrony really fast.” Elowitz is part of a burgeoning field called sys- Perhaps this shouldn’t have come as a surprise, tems biology, an outgrowth of molecular biology, since biology—and the life of a cell—is known to which has mainly focused on the roles of individual be complicated and messy. Gene expression—the molecules and their structures. Systems biologists, process by which a gene turns on to make a however, seek a broader understanding of how protein—is itself a muddled affair. Consisting of these parts underlie the mechanics of life. Tradi- many biochemical reactions, gene expression is, in tionally, research papers would propose a series of essence, a mechanical dance. Although choreo- interactions to explain a particular phenomenon, graphed well, it may be ungainly and missteps are according to Elowitz. But this frustrated him, he inevitable. Molecules called transcription factors says, because papers stopped short of quantify- bind to specific DNA sequences and help genes ing the important components. Imagine some- get made into proteins. But they may not bind one giving you a cake recipe that told you to use properly, or they may fall off the strand before their flour, eggs, milk, and butter, but didn’t tell you job is done. Some of the molecules involved may how much of each you needed. Given the inher- number fewer than 100—or even 10—making ent complexities of genetic circuits and biological these missteps even more pronounced. As a result, systems in general, studying them in detail is a a genetic circuit doesn’t work the exact same way challenge. While a graduate student at Princeton, every time. There’s a certain amount of inherent Elowitz decided to approach the problem from noise. Like the random variations that cause static another direction—he set out to make his own in your AM radio signal, this noise was likely the simple circuit. After all, what better way to learn culprit that knocked the represillators out of sync. how to bake a cake than to try your own recipe? Biologists have long suspected that cells and their In 2000, Elowitz and his thesis advisor, Stanislas internal biochemistry might be inherently noisy. Leibler, designed and built a simple biological clock But in a 2002 paper, Elowitz, then at Rockefeller based on proteins called repressors, which turn University, and colleagues were able not only to off specific genes. As with all proteins, repressors identify and quantify noise in cells, but to separate are made when their genes are turned on, so the it into two components they called intrinsic and researchers engineered a set of three genes that extrinsic noise. Extrinsic noise is due to fluctua- turned each other on and off, like a three-way tions throughout an individual cell—differing rock-paper-scissors game. In such a loop, rock concentrations of the transcription factor, for turns off scissors. But scissors turns off paper, so instance. Intrinsic noise comes from the process of if scissors is off, paper remains on. Paper, on the gene expression itself. To isolate these two kinds other hand, turns off rock, so if paper is on, rock of noise, the researchers put two nearly identical turns off. This then causes scissors to turn back genes into the same cell. The researchers engi- on, and the entire sequence cycles through as each neered each gene to make a different color-coded

2 0 0 8 ENGINEERING & SCIENCE NO . 1 29 This series of time-lapse frames shows the dynamics of cell behavior over the course of nearly two-and-a-half days. The cells that are expressing ComS glow green, and the white shapes are spores. After the cells have divided some, one cell begins to glow red, indicating competence. During competence, the cell can’t divide, although it still grows. Only when the cell leaves its competent state is it able to divide again.

fluorescent protein, so that they could track the IT’S JUST LIKE FLUSHING A TOILET gene’s expression. This powerful technique, now commonplace, was only made possible when Commonly found in soil, B. subtilis is sometimes biologists cloned fluorescent protein genes from used to make a type of fermented soybean eaten in jellyfish off the west coast of North America more Korea and Japan. But Elowitz and his colleagues than a decade ago. Although many scientists have were interested in what B. subtilis did when there used fluorescent proteins and movies to study cells, were no soybeans to ferment. When faced with the Elowitz Lab is among the first to use them stress—such as a lack of nutrients—B. subtilis in a highly quantitative fashion. Much like how sometimes assumes a state called competence. In oscilloscopes allow electrical engineers to measure its competent state, B. subtilis can take in stray an electrical circuit, these movies allow biologists to strands of DNA that happen to be floating around. measure a genetic circuit, Elowitz says. Many scientists say competence is bacterial sex, a If the noise were mainly extrinsic, it would means for it to exchange genetic material. Others affect the genes in the same way, and the research- contend that it allows B. subtilis to eat the DNA as ers would see equal amounts of each color in any food. particular cell. For example, if the two colors were But what interested Elowitz was that compe- red and green, then every E. coli cell would appear tence, like many other types of cellular differentia- a uniform yellow. Alternatively, if the noise were tion, is not a predetermined fate. Instead, each cell mainly intrinsic, then each gene would behave has a certain probability of becoming competent or differently. You would now expect intermediate not. Even when they have the same genes and are shades such as orange-yellow and goldenrod as put in the same environment, only a few percent well as red and green cells. The cells behaved as of the cells ever become competent at a given time. the team had hoped. The researchers engineered This is good for the bacteria, becauseB. subtilis a strain dominated by intrinsic noise, and the can’t divide when competent. If all became com- cells glowed a whole spectrum from red to green. petent at once, then none could divide, spelling Likewise, for colonies with minimal noise, the cells doom for the colony. The researchers also discov- were all yellow. The researchers also found that ered that the competent state is temporary. After cells were more susceptible to intrinsic noise when less than a day, a competent cell returns to normal. genes were expressed at low levels. “Noise is not But how do these cells, which are genetically iden- just some mysterious fluctuation that causes cells tical, know which ones should become competent, to be different from each other,” Elowitz says. “We and how do they know when to turn back? “It’s can understand why cells are different from each very mysterious,” Elowitz says. “All the cells are The fluorescent proteins in other, and we can break [noise] into these different supposed to be the same, but they all do something components.” different, even when you control the system as this image track sporula- These direct measurements of noise demonstrat- carefully as possible.” tion in B. subtilis. Cells ed its importance, but what wasn’t clear was how The key player is a transcription factor called about to sporulate turn noise influenced a genetic circuit and, subsequent- ComK. ComK activates the genes needed for red, while the part that’s ly, cellular behavior. The question was twofold: competence as well as the gene that makes ComK how does a cell suppress fluctuation when it wants itself, which is written as comK, forming a positive about to become the spore to do something accurately, and how can a cell use feedback loop. A second player is a protein called turns yellow. the noise to its advantage? The answer to the latter, ComS, made when one cell sends another a chemi- at least, lay in another bacterium: B. subtilis. cal signal; the reception of that signal activates

30 ENGINEERING & SCIENCE NO . 1 2 0 0 8 “What’s great about this,” Elowitz says, “is that we went from the repressilator, where noise was an annoyance—we found it wasn’t working precisely because of the noise—to a natural differentiation circuit where the cell has actually taken advantage of this seeming annoyance to produce a system that can control the probability of whether something happens.” To develop their model, the researchers used fluorescent proteins that glowed different colors when ComK and ComS were expressed. The brightness of each color corresponded to how much ComK and ComS was being made, with the glow of ComK heralding competence. They then took snapshots of the bacteria colonies every 20 minutes Diagram of the basic com- or so, allowing them to track each cell. petence circuit. Their analysis showed that the competence The key player is ComK, circuit was an “excitable system,” in which random fluctuations trigger some process after crossing a the protein that trig- particular threshold. A simple example of an excit- gers competence. ComK, able system is jiggling the handle of a toilet. “Once which also promotes in a while, you’re going to jiggle a little too hard, itself, as indicated by the and it’ll just cross the threshold, and it’ll flush,” Elowitz explains. Another characteristic of an dotted arrow, blocks the excitable system is that the strength of the trigger expression of ComS, as doesn’t affect the system. Pushing the toilet handle indicated by the T-shaped harder doesn’t make the flush more powerful or arrow. The third player is last longer. In the competence circuit, the expression of ComK is inherently noisy—like jiggling a called the MecA complex, the gene—known as comS—that makes ComS. toilet handle—and once in a while, enough will be an enzyme that destroys Finally, a third party called the MecA complex joins expressed to trigger competence. But more ComK both ComK and ComS. the dance. MecA causes a protease, an enzyme that doesn’t cause the cell to become competent more destroys proteins, to target both ComK and ComS. often or for a longer time. When enough ComS is Similar to the repressilator, the three proteins The competence circuit is a self-controlling one. produced, MecA goes after work together in a network of negative and posi- Once the system turns on, it sets up a chain of ComS instead of ComK, tive feedback loops. By destroying ComK, MecA events that eventually turns itself back off. “The allowing enough expres- blocks competence. But when enough ComS is system can spontaneously flick on, but when it’s expressed, MecA becomes occupied zapping ComS on, it starts to activate something that will build up sion of ComK to start instead of ComK. The production of ComS, then, and build up,” Elowitz explains. “When it goes to a competence. In this way, indirectly supports competence by acting as a decoy high enough level, it’ll shut the whole thing down.” ComS prevents MecA from for MecA. It also turns out that, through a series of The competent state happens while MecA is busy blocking ComK. intermediate interactions, ComK blocks the expres- gobbling up ComS, and additional ComS isn’t sion of ComS, which prevents any more decoys being made because ComK is repressing it. Over from distracting MecA. MecA then can attack time, ComS dwindles—in this case, what’s “build- ComK, eventually shutting down competence. ing up” is the depletion of ComS. Once MecA Although past research had unearthed this batch runs out of ComS to destroy, it goes after ComK. of interactions, no one had been able to explain Since ComK is the trigger for competence, repress- the subtle interplay between each step. Here was ing ComK returns the cell to its normal state. another case of a cake recipe consisting only of a This work is among the first to show how cells list of ingredients. In 2006, however, then-postdoc can use noise to their advantage. Evolution has Süel, along with Elowitz, graduate student Louisa created an excitable system that allows just the right Liberman, and Jordi Garcia-Ojalvo from the Tech- fraction of B. subtilis cells to become competent. nical University of Catalonia in Spain, were able to But even though the model is consistent with what mathematically model the competence dance. The B. subtilis does, how do you know for sure that model didn’t get bogged down in every single bio- noise is indeed the driving force behind differentia- chemical reaction, but captured the essence of the tion? Elowitz points out that the decision mak- genetic circuit. It showed that the cell’s decision ing could be hiding in an unknown interaction. to differentiate relied on random noise—the same The researchers continued tweaking and tuning kind of fluctuations that prevented the repressilator the genetic circuit, and in an elegant experiment from pulsing in sync among all the cells. In other showed that without noise, competence was not words, to determine which cell would become possible, Elowitz says. “This was kind of the proof competent, B. subtilis drew straws. that noise was necessary.”

2 0 0 8 ENGINEERING & SCIENCE NO . 1 31 Left: As the size of the cells increase, the probability of their becoming competent decreases, as predicted by the model. The experimental results

(black line) are consistent 2007. 23, March 1718, p. 315, vol. ,

with computer simulations Science (gray line). Far right: An image of the

extra-long B. subtilis. The al., et Süel From red cells are competent.

TUNING OUT THE NOISE won’t flush the toilet if you barely jiggle the handle. In fact, the probability of a cell becoming compe- Noise is all about numbers. If you want to know tent went from about 3 percent to less than a half America’s favorite ice cream flavors, but only ask 10 percent as the cell lengthened. people, you probably won’t have much confidence The team also discovered that the minimum, or that your results are representative of the general basal, expression level of the comK gene controls population. But if you ask 100,000 people, the the frequency of competence, and the basal expres- results will surely be more accurate. With more sion level of the comS gene independently controls numbers, the uncertainty—i.e., the noise—goes the duration of competence. In the context of the down. Noise is prevalent in the cell because cells model, this makes sense, since the amount of the are tiny, and molecules like ComK number in only ComS protein acts like a timer in the circuit; the the tens or hundreds. To boost the numbers, Süel, cell is competent until MecA finishes off ComS graduate student Rajan Kulkarni (PhD ’06) , Jona- and turns to ComK. Likewise, ComK controls the than Dworkin of Columbia University, Garcia- decision to enter competence in the first place. Ojalvo, and Elowitz reduced noise by engineering By tweaking the two basal expression levels, the versions of B. subtilis that were up to nine times researchers tuned B. subtilis to become competent their original lengths, and thus contained nine more often or for a longer time. Increasing the times as many molecules like ComK. basal expression level for ComK by about 20 times To make the extra-long bacteria, the research- its normal amount caused all cells to become com- ers introduced a mutation that prevented the cells petent. By increasing the basal expression level for from fully dividing. Now, during division, each ComS, the team was able to make most of the cells cell would double in remain competent for about 40 hours, compared size, but not split off. with a normal duration of about 20 hours. Since the molecu- This finding is more than a bioengineering nov- lar concentrations elty, however. The fact that competence frequency remained the same and duration can be tuned independently may be throughout, with big- critical for a bacterial species’ evolutionary survival. ger cells came more Scientists still don’t fully understand the reasons molecules and there- behind competence, but if the reasons were to fore less noise. The enhance genetic diversity, then competence would researchers’ model allow B. subtilis to evolve a response against a stress- predicted that with ful environment. Say that the time the organism reduced noise, the cells spends in competence is already evolutionarily fine- would not become tuned. If a more stressful environment favors more competent. Fewer B. subtilis to become competent, then the organism fluctuations meant a can evolve in that direction without affecting the smaller chance that optimized duration. According to Süel, it’s like enough ComK would designing a race car: “Maybe you want to improve A group of green-glowing B. subtilis, signifying be expressed to trigger the braking performance but don’t want to sacrifice that ComS is being expressed. A few cells have differentiation. To the steering.” become white spores. return to the toilet The team then wanted to see what would hap- analogy, you probably pen if they added another feedback loop to the

32 ENGINEERING & SCIENCE NO . 1 2 0 0 8 competence circuit. They inserted a protein called the details are different: the organism has more Rok that represses the expression of ComK. They genes or different interactions. But the approach, anticipated that the addition would spur the cells the thinking, and the tools we’re using are not to leave competence faster, and indeed that’s what limited to this one particular bacteria.” happened. Instead of 20 hours, most of the cells In the past, biologists were primarily interested were competent for only about 14 hours. But then in looking at whether specific genes are expressed the modified circuit offered a surprise: there was or not, the researchers say. Now, scientists want to less variability among the cells than in the normal learn about the dynamics of gene expression—how circuit. Remember that since every system has fast does the gene express itself, and at what level? a certain amount of noise, every measurement How does it change over time? features an inherent variability. Even though most According to Elowitz, researchers now have of the normal cells were competent for about 20 enough biochemical data about genetic interac- hours, a few would remain in that state for 10 or tions to develop mathematical models that can 30 hours. In the new circuit, more cells stayed make quantitative predictions. Elowitz, trained competent for similar amounts of time. in physics, winnows complexity into simplic- ity. A complicated network of interactions, the competence circuit, has been distilled into two equations describing the dynamics of ComK and “What’s important is the creativity of the experiment. Can we make simple, ComS. “Sometimes a physics point of view lets you ask questions you otherwise wouldn’t ask,” says beautiful, and elegant experiments that push this field into new directions, or Elowitz. Perhaps it was this kind of interdisciplinary allow us to see things in a simpler, clearer way?” thinking that earned him a MacArthur Fellow- ship, which, as its website says, is given to “talented individuals who have shown extraordinary original- ity and dedication in their creative pursuits and a marked capacity for self-direction.” The distinction The researchers were able to explain this includes $500,000, given over five years, for the enhanced precision with their model. Normally, individual to do with as he or she pleases. when a competent cell returns to its normal state, The recognition, he says, was a nice pat on the there are few ComS molecules, since its depletion back. “It’s encouraging,” he says. “It makes you is what allows MecA to gobble up ComK and send stop and think about what’s important, and what’s the cell back to normal. But with the addition important is the creativity of the experiment. Can of Rok to help block ComK, the cell returns to we make simple, beautiful, and elegant experiments its normal state earlier, while there are still many that push this field into new directions, or allow ComS molecules left. Since ComS governs compe- us to see things in a simpler, clearer way? It makes tence duration, and the reengineered circuit has left you ask yourself, are we doing enough? Are we a higher number of ComS molecules, there is less doing the right things? Are we doing things that noise in the distribution of duration times. “This is are too easy?” He says he hasn’t decided what to do weird for biological circuits,” Elowitz says. “Nor- with the money yet, but the fellowship nevertheless mally, when you mess around with these genetic propels him forward. “People think what we did circuits, you’re going to screw them up in some so far is good,” he says. “We have to do something way.” crazier now.”

TOWARD SOMETHING CRAZIER?

The ability to understand a genetic circuit at such a deep level is an accomplishment in and of itself, but this work has broader implications. Many other single-celled organisms, not to mention cells in multicellular organisms, diﬀerentiate on a probabilistic basis. Additionally, being an excitable system is not unique to the competence circuit. Biological systems at scales other than the gene level are also excitable—the ﬁring of neurons in the brain, for example. “You have a tree, a rat, or a human—they’re all governed by proteins, DNA, and RNA,” Süel says. From left to right: Gürol Süel, Michael Elowitz, and Jordi “The rules for how these molecules work together Garcia-Ojalvo. and give physiological behavior at the cellular level seem to be conserved among all organisms. Maybe PICTURE CREDITS: 28-33 — Elowitz lab; 31 — Doug Cummings

2 0 0 8 ENGINEERING & SCIENCE NO . 1 33 Ad Astra per GALCIT

By Douglas L. Smith

Just about every kid wants to be an astronaut at some point. TheStar Wars and Star Trek franchises are still going strong 30 and 40 years after their inception, and the commercial satellite business brought in some $82 billion dollars last year worldwide. Amazing pictures, from Mars to distant galaxies, are only a few mouse clicks away. Yet U.S. aerospace firms can’t fill some 10 percent of their MS- and PhD-level engineering jobs, according to industry insiders. The problem will just get worse as the rest of the bright young minds who came of age in the years of the Apollo program retire—28 percent of them will have called it a career by the end of this year, says the National Science Board, which is the governing body of the National Sci- ence Foundation. It’s not that kids get bored with space, says Ares Rosakis, the von Kármán Professor of Aeronautics and Mechanical Engineering and director of the Graduate Aeronautical Laboratories at the Califor- nia Institute of Technology (GALCIT). “Space is actually the easiest way to start teaching science in ages below 16.” The problem, he says, comes when those 16-year-olds start thinking like grownups— when they start worrying about a career. “I see it in my own children. They have been indoctrinated since they were young about how fantastic science Rendering courtesy of John Friedman Alice Kimm Architects. Kimm Alice Friedman John of courtesy Rendering is, but their classmates tell them it’s not cool to become an engineer—not only because it’s difficult, Air flowing over and around objects in its path can create trains of spiraling eddies known but because my father makes much more money as von Kármán vortices, as seen (top and right) over the island of Rishiri in the northern than yours.” College graduates with engineering Sea of Japan. In Caltech’s remodeled Guggenheim Aeronautical Laboratory (above), swirling degrees tend to go into information technology, biotech, or some other -tech, or even farther afield shapes resembling von Kármán vortices will decorate the ceiling of the conference room to business, law, or investment careers that seduce named for the man who first described them—aeronautics professor and GALCIT founding them with promises of millions before they’re 30. director Theodore von Kármán. If there’s any place that can make space engineering cool again, it’s got to be Caltech. After all, the Institute founded and still administers NASA’s Jet Propulsion Laboratory, which got America into the Space Race and is now exploring our solar system and beyond. Since 2004, Rosakis has been spearheading an initiative to launch a master’s

34 ENGINEERING & SCIENCE NO . 1 2 0 0 8 degree program at GALCIT that would go beyond lineup includes four offerings: the traditional MS in aeronautics, which keeps Ae 115, Spacecraft Navigation, covers astrody- one safely within Earth’s atmosphere. The new namics, orbital calculations, and precision track- MS in aerospace engineering “reconnects to JPL ing systems. The class is being taught by Michael in a major way,” says Rosakis. “It engages JPL, Watkins, who often draws examples from his Caltech’s seventh division, in curriculum design work as the mission manager for the Mars Science and classroom teaching. I don’t think I would dare Laboratory (MSL), the “super rover” set to launch do a space-engineering master’s program anywhere in 2009, and as the project scientist for the Grav- else that didn’t have JPL next to it. This program ity Recovery and Climate Experiment (GRACE), is illuminated by JPL’s existence.” In the program, launched in 2002. MSL requires ultra-precise students work on a real spacecraft, solving real navigation to enter the Martian atmosphere at engineering problems. “We are not designing exactly the right time and place for a safe landing. entire missions. You could imagine us trying to GRACE consists of two satellites orbiting Earth launch a Mickey Mouse GALCIT experiment. in tandem about 220 kilometers apart, with their It’s much better to isolate a small part of a real relative positions determined to an accuracy of less mission.” Rosakis credits JPL director Charles than one micron—akin to measuring the distance Elachi (MS ’69, PhD ’71) for this approach. “His between JPL and San Diego to within the thick- philosophy is to expose the students to some- ness of a red blood cell. That knowledge allows the thing of real importance to JPL that would allow mapping of minute changes in Earth’s gravitational JPL instructors to be enthusiastic because they field due to climate-related effects. “We even are working, through teaching, on an actual JPL weigh Earth’s ice sheets,” says Watkins. mission.” But although the JPL instructors are Ae 121, Space Propulsion, looks at solid- and working in their areas of expertise, they are paid liquid-fueled rockets and their control, plus such by Caltech, not JPL. “This is very important,” exotica as electrical and even nuclear thrusters. Rosakis emphasizes. “The JPLers are employed as This course features James Polk, principal inves- consultants, as instructors separately from their JPL tigator for the NEXIS (Nuclear Electric Xenon duties. It’s against NASA rules to use government Ion System) engine, a high-power electric thruster contracts for teaching.” designed to propel a nuclear-powered spacecraft to The one-year aerospace program, now in its the icy moons of Jupiter. NEXIS was an advanced second year, is just as big as the aeronautics pro- version of the ion drive demonstrated on JPL’s gram—nine students each per year—and, like the Deep Space 1, which flew by comet Borrelly in traditional degree, is designed to be a stepping- September 2001 and took the best-ever pictures of stone to a PhD. Eight of the students from the a comet’s nucleus. inaugural class have, in fact, gone on to Caltech Ae 159, Space Optical System Engineering, PhD programs in space-related fields. Because includes not just sensors, lenses, and mirrors but the program is not intended to lead to a terminal all the other considerations needed to bring the master’s, it focuses on the science behind space universe into focus: the thermal properties of engineering rather than on job-specific training. materials, the design of servomechanisms, and the “It’s not an applications-based program,” says subtleties of image processing. The instructor is Rosakis. “That’s a very big difference from the rest James Breckinridge, JPL’s Origins theme technolo- of the space graduate programs you see around the gist, which is a fancy way of saying he oversees tele- country.” Another big difference is the overkill scope technology at the Lab. Breckinridge headed factor. Four GALCIT members—fully one-third the nationwide team that gave an eye exam to the of the aeronautics faculty—and six JPLers teach the flawed mirror on the Hubble Space Telescope. He nine students. The basics of fluid mechanics, solid also led the JPL team that built the Wide-Field and mechanics, and structural mechanics are addressed Planetary Camera 2 to prescription to compensate. in all their gory detail in a set of core courses com- And EE/Ge157, Introduction to the Physics of mon to both the aeronautics and the aerospace Remote Sensing, covers devices across the electro- option. The math is just as strong as it would be magnetic spectrum, from microwave radars that for a PhD, “and so of course our students have look beneath the sands of the Sahara to infrared suffered,” laughs Rosakis. “But that’s life.” But the spectrometers that take the temperatures of distant pain pays off when a prepared mind finds the right planet-forming systems. This very popular class, situation, as Simon Ramo (PhD ’36) did when he originated and taught by synthetic-aperture radar went to work for General Electric armed with all scientist Charles Elachi long before he became the latest physics that was about to revolutionize JPL’s director, is now led by Jakob van Zyl (MS electrical engineering—see the sidebar on page 17. ’83, PhD ’86), an Elachi protégé who started off The program’s elective courses are taught by JPL flying Earth-mapping radar missions on the Space staff members. This is overseen by JPL’s chief tech- Shuttle and now leads JPL’s Astronomy and Physics nologist, Paul Dimotakis (BS ’68, MS ’69, PhD Directorate. ’73), who is also Caltech’s Northrop Professor of But the plum course is a required one—Ae 105, Aeronautics and professor of applied physics. The Aerospace Engineering, co-taught by Dimotakis, exact slate varies from year to year, but the current Gregory Davis, and Marco Quadrelli. Dimota-

2 0 0 8 ENGINEERING & SCIENCE NO . 1 35 kis is an expert in turbulent flows in general and see the hardware design and procurement. The propulsion systems in particular. Davis is the chief third year, 2010–11, will be NuSTAR’s assembly, technologist for JPL’s Mechanical Systems Division; test, and launch. And the final year, 2011–12, will before that he built the advanced deployable struc- monitor the hardware’s performance in orbit. “We tures group, and before that he was a mechanical- will repeat this cycle every four years with a new systems engineer on the Mars Exploration Rovers, JPL mission,” Rosakis says. Spirit and Opportunity, and a dynamicist for the Next year, Ae 105 will be overseen by Professor original Sojourner rover. Quadrelli is a senior of Aeronautics and Civil Engineering Sergio Pel- dynamics and control analyst in JPL’s guidance, legrino, just recruited from the University of Cam- bridge, England, where he founded the university’s Deployable Structures Laboratory in 1990. (He is also teaching Ae 221, Space Structures.) “There are many joint Caltech-JPL appointments,” says “One cannot change the world with nine students, but perhaps one can. Rosakis, “but they took the joint appointment after being hired by Caltech. This is our first professor I would be very happy if one of them became a future director of JPL. . . and to be hired by a joint Caltech-JPL committee.” maybe in 10 years one will become another von Kármán.” This arrangement will give JPL access to a large chunk of Pellegrino’s time, and allow him unfet- tered access to JPL. Pellegrino is no stranger to expandable structures. His office is crammed—or will be, when he navigation, and control section who has been an finishes unpacking—with models that would make entry, descent, and landing dynamics analyst for excellent toys, should someone care to mass-pro- the Huygens probe to Saturn’s moon Titan and the duce and market them. Some of them are solid Mars Exploration Rover and MSL missions, and shapes, like the circular plate that expands into has developed control algorithms for multivehicle a lacy circle about twice its original size. Others and tethered spacecraft. Ae 105’s fall term cov- look like bundles of sticks, joined by hinges on ers launch vehicles, fundamental structural and their ends. Shake out the bundle, and four sticks thermal systems, and introductory orbital mechan- become a square, six a hexagon, and so on. Put ics. The winter quarter looks at advanced orbital a wire mesh on that framework, says Pellegrino, mechanics, spacecraft dynamics, and planetary and voilà! You have a radar array. Pellegrino and reentry—taking the students from the ground to his collaborators have worked out general math- space and back again in two terms. Occasional ematical techniques for dealing with both classes guest lecturers from JPL provide additional spice to of objects, so that “when people have a particular this challenging stew. problem, they can go to the general theory and pull In the spring term, the Ae 105 students reach out the specific geometry they need. I don’t like to the promised land of flight hardware. The class is work on problems no one could make use of.” doing a theoretical analysis of the telescoping mast In the years to come, Pellegrino and Harrison for the Nuclear Spectroscopic Telescope Array, or will be co-teaching a course devoted to NuSTAR. NuSTAR, slated to launch in 2011. A partnership It’s not yet in the catalog, and so doesn’t have an between Caltech, JPL, and 10 other organiza- tions and headed by principal investigator Fiona Harrison, professor of physics and astronomy, NuSTAR is an X-ray telescope. Since X rays go through pretty much everything, you have to place any optics intended to focus them at a very shallow angle. This means they have to be quite some distance from the detector, which is a problem if you’re designing a compact spacecraft that’s relatively inexpensive to launch. Hence the deployable mast—essentially a set of collapsible cubes stacked one atop another—which, when fully extended, will hold the mirrors 10 meters away from the spacecraft. This summer, some of the Ae 105 students and some JPL new hires assigned to the mission will begin actual deployment experiments with real prototypes. In the fall, the 2008–09 academic year will mark the start of a four-year program of col- Pellegrino holds an expandable sphere—a 3-D version of laboration with the NuSTAR team as the mission develops. That first year, students will be doing the expandable circle mentioned above—that might make mission planning. The second year, 2009–10, will a dandy space habitat as a hollowed-out shell.

36 ENGINEERING & SCIENCE NO . 1 2 0 0 8 GALCIT-JPL connections then and now. Right: Allen Puckett (PhD ’49), a grad student of von Kármán’s, designed several supersonic wind tunnels for the then-Army laboratory. In the front row at the January 1947 ground- breaking for one of them are, from left, Clark Mil- likan (PhD ’28), professor of aeronautics and future GALCIT director; Puckett; official number, but it will be called On-Orbit Aeronautics and professor of bioengineering; Assis- Louis Dunn (BS ’36, MS ’37, Performance of Large Deployable Space Structures. tant Professor of Aeronautics Beverley McKeon; “As the design and construction of the mission and option administrator Dimity Nelson] was very MS ’38, PhD ’40), JPL direc- continues, the students will be doing hands-on forward-looking.” The interior’s organic forms tor and associate professor work with the flight hardware,” says Pellegrino. “were based on turbulent flow diagrams from the of aeronautics; Caltech They’ll be creating a diagnostic package that will aerospace faculty,” says Friedman. “Now you don’t use cameras and other techniques to watch the send mechanical drawings to the shop any more, president Lee DuBridge; boom as it deploys. Rosakis says the package will you send digital files to a CNC [computer numeri- Major General Everett S. “look at the mechanics of the expanding boom, cal controlled] machine.” Hughes; von Kármán; and and make sure that it goes where it has to go and All the corridor walls follow graceful curves Robert A. Millikan, profes- how it vibrates in the process. That has nothing to rather than straight lines. “There are no oblongs do with the science mission of NuSTAR directly. here. No circles, either. It’s all about the flow,” says sor of physics and Caltech It has to do indirectly, because it helps in the posi- Rosakis proudly. The lobby features a translucent, president emeritus. tioning of the instrument, but it is a mechanics and undulating ceiling with large, shallow, backlit dim- Far right: Two of the an optics problem.” ples that produce moiré-like patterns. The lantern, aerospace master’s degree The grad students won’t have all the fun—there as Kimm calls it, is reminiscent of the dimpled will also be Summer Undergraduate Research wing surfaces that McKeon studies. Decorative program’s prime mov- Fellowships (SURF) available. “Fiona was very elements throughout the building reflect other ers—Rosakis (at far right) inspired by President Chameau’s call for new researchers’ work, and a number of artifacts from and Dimotakis (at far opportunities for the students to learn,” Pellegrino the old building have been incorporated as well, says. including a control panel from the old 10-foot left)—are alumni of Athens The new program is housed in a completely wind tunnel that has been mounted into a wall on College, a Greek-American remodeled Guggenheim Aeronautical Laboratory. a third-floor corridor. high school in Athens, Built in the late 1920s as a shell wrapped around The centerpiece of the new Guggenheim—the Greece . . . as are (middle a wind tunnel with a 10-foot-diameter working Laboratory for Large Space Structures—is Pel- section—the biggest and fastest wind tunnel on the legrino’s province. The lab occupies much of the left) Sterge Demetriades West Coast, capable of generating sustained flows ground floor and includes a two-story-tall “high (Eng ’58), retired from a of up to 200 miles per hour—much of the cavern- bay,” an open space big enough to accommodate long aerospace career, and ous space has been underutilized since the wind an Apollo moon lander. Windows look into the Alexis Livanos (BS ’70, MS tunnel was decommissioned in 1997. The new high bay from the second floor outside, making it interior, designed by John Friedman Alice Kimm part lab and part interactive display. The high bay’s ’73, PhD ’75), president of Architects, has already won a 2007 American Insti- first occupant will be NuSTAR’s truss—a per- Northrop Grumman Space tute of Architects, Los Angeles Chapter, Next L.A. fectly rigid structure in the microgravity of space, Technology and chair of Merit Award—one of 10 among a couple of hun- but incapable of supporting its own weight when dred submissions. “This project allowed us to do a extended to full length parallel to the ground. In the GALCIT advisory coun- little more experimentation than usual in terms of order to test such ultralightweight—one could even cil. This picture was taken design,” says Kimm, the lead architect. “It was a call them flimsy, but in a good way—structures at the 50 Years in Space really exciting project, with lots of digital model- here on Earth, they need to be “gravity compen- conference. ing techniques. The design committee [Rosakis; sated,” which essentially involves suspending them Guruswami Ravichandran, the Goode Professor of from wires like X-wing fighters hanging from the Aeronautics and Mechanical Engineering; Morteza ceiling of a 14-year-old’s bedroom. The suspension Gharib (PhD ’83), the Liepmann Professor of systems may include pulleys, counterweights, and

2 0 0 8 ENGINEERING & SCIENCE NO . 1 37 servomotors to balance the changing forces. “As ’43, PhD ’46) are all GALCIT alumni. Two of the structure deploys, the suspension has to move them have left quite a mark on aerospace them- with it,” Pellegrino explains. “And so a high ceiling selves. Puckett, now retired as chairman of Hughes is good, because the angles of the wires change Aircraft, designed and built the United States’ first less.” Mounting the pulleys on tracks doesn’t help, supersonic wind tunnel while still a grad student. says Pellegrino. “Fixed wires are better than tracks, And Charyk, a former undersecretary of the Air because the structure would only be able to go in Force, was the founding president of the Com- a straight line to follow the track. And tracks have munications Satellite Corporation, or COMSAT, friction, so they’ll drag. So wires are actually better, which in 1965 launched the world’s first geosyn- not to mention cheaper!” Pellegrino also works chronous communications satellite to beam TV with inflatable structures—if you’ve ever blown up broadcasts and telephone conversations between an air mattress straight out of the box, you know Europe and North America. that it doesn’t always unfold precisely as planned— Even though the new labs will have been in and with composite materials that can be formed use over the summer, they will officially open on into thin, stiff things like antenna dishes. Equip- September 26, when President Chameau cuts ment for making and testing these kinds of struc- the ribbon as part of the celebration of the 80th tures will be housed in the lab as well. anniversary of Guggenheim’s completion and Wrapped around the second floor of the high the birth of GALCIT. But the highlight of this bay is the Gordon Cann Laboratory of Experi- two-day party will probably have been the previ- mental Innovation. This lab will house Ae/APh ous evening, when astronaut Gregory Chamitoff 104, Experimental Methods, which, says Rosakis, (MS ’85) addresses the banquet attendees from “is a signature course. No other university has aboard the International Space Station just before a required full-year graduate-level experimental GALCIT’s Aerospace Historical Society awards the course. They are all simulations and numer- 23rd Annual International Wings of von Kármán ics.” The class teaches solid- and fluid-mechanics Award to Alexis Livanos (BS ’70, MS ’73, PhD methods of all sorts in one dedicated lab space. ’75), president of Northrop Grumman Space Tech- The facility includes clean rooms, a transmission nology and a 2008 Caltech Distinguished Alum- electron microscope, materials testing equip- nus. Incidentally, that afternoon’s keynote speaker, ment such as a coherent gradient sensing (CGS) David Thompson (MS ’78), chairman and CEO of laser-interferometry system for measuring how Orbital Sciences Corporation, once said, according materials and structures deform under stress, and to Rosakis, that “the reason he left GALCIT, as did particle-velocimetry equipment to see how fluid Chamitoff—they did not get PhDs with us—is flows behave, and even small wind tunnels. “Stu- that we did not have a concentration in space. This dents used to have to go from lab to lab and from shows the need. Thompson actually said that that professor to professor for all these different things,” was one of his biggest disappointments, because he Rosakis says. Down the hall will be the new von really wanted to work in space.” Kármán Conference and GALCIT Archives room, And work in space these new degree holders will. underwritten by Bob Herzog. The third floor will Virendra Sarohia (MS ’71, PhD ’75), who works include two new labs for student experiments, and in the chief technologist’s office at JPL as liaison to the Allen Puckett Laboratory of Computational Caltech, is in charge of making that happen. He’s Fluid Mechanics, which includes a seminar room on the Caltech-JPL committee that recruits JPLers and open-plan computer-lab area, plus open spaces to teach in the program, but his broader role, says designed to hang out in. Rosakis, is to facilitate the research connections The renovations have even spilled over into the between Caltech professors and JPL scientists in a second floor of the adjoining Karman Laboratory way that will lead to thesis topics for the students. of Fluid Mechanics and Jet Propulsion, where the “So that’s why it’s important that he is a GAL- Joe Charyk Biomechanics Laboratory has been CIT graduate who is intimately familiar with JPL located. This facility will house work being done research and Caltech research.” Besides brokering by Gharib and by John Dabiri (MS ’03, PhD ’05), these “arranged marriages,” Sarohia oversees the assistant professor of aeronautics and bioengi- JPL summer fellowship program that lets program neering, on bioinspired designs of such things as graduates work full-time in space before going on low-power propulsion systems for exploring worlds to a PhD and, Rosakis hopes, a career in the space with dense atmospheres—picture a robotic jellyfish business. “One cannot change the world with pulsating its way through Jupiter’s clouds, for nine students, but perhaps one can. I would be example. (For more on Dabiri’s work, see his 2007 very happy if one of them became a future director Caltech News profile, in Volume 41, Number 1.) of JPL, and a few others become CEOs of major Cann (MS ’56, PhD ’61), Herzog (BS ’56, MS aerospace companies in Southern California, and ’63, Eng ’64), Puckett (PhD ’49), and Charyk (MS maybe in 10 years one will become another von Kármán.”

Von Kármán vortices off the Chilean coast near the Juan Fernández Islands. PICTURE CREDITS: 36, 37 — Bob Paz; 34–35 — NASA Landsat; 37 — Caltech Archives

38 ENGINEERING & SCIENCE NO . 1 2 0 0 8 Letters

Thank you for the special The article by Erik Con- issue of E&S on JPL. It way, arriving shortly after my read like a very good novel. 90th birthday, really stirred Even though I was familiar memories. with all of the missions that I would have been the class you described, the details of of ’39, but took a year off the problems that had been to work for the New York addressed were fascinating. World’s Fair of 1939, followed My familiarity with JPL by returning for my senior as a Caltech student, hav- year with my new bride, a ing classmates on the staff, Hungarian girl who’s fluent working with them in setting in about five languages. This up Seminar Day talks, and put me in with a group of following the adventures in postgraduates, all delightfully the news media gave particu- somewhat weird, including lar significance to your story. Frank Malina. As I followed the pictures I Part of our association was recognized many as persons because I was a good mechan- that I had met or worked with ic and there were things in industry. to be performed on rocket My one compelling thought motors, and mostly because was that it would be great if my wife was good at translat- some sort of recognition could ing various technical papers be afforded to JPL itself. I for this group and some other doubt that the ground rules postgraduates. for the Nobel Prize could be Your article led me to dig stretched far enough to permit out my scrapbook, made their inclusion in such an from 35-mm film that has award, but I do feel that their long since deteriorated, but I contribution to the explora- enclose not only the original tion of the Solar System prints but some enlargements should receive some sort of from my office copier that international recognition. show one of the firings, sand- Thanks again for your work. bags and all. The group played when Alfred W. Thiele (BS ’51) they weren’t rocketing, and I According to Cox, these old scrapbook photos show have other photographs show- Frank Malina testing rockets—apparently in someone’s ing us camping on the Mojave backyard. If anybody knows where these photographs and trying to shoot jackrab- were taken (Cox doesn’t remember), we’d love to hear bits with a .45. I graduated and went to from you! I thoroughly enjoyed your work for Menasco airplane “50 Years of Caltech and JPL” engines, and all of a sudden issue. Very well written. An JATO came along and the excellent selection of material. weird but wonderful group all As an alumnus who closely became VIPs. followed the many spectacular events chronicled, it was great Robert O. Cox (BS ’40) E&S welcomes letters. Send correspondence to to get such a well-done com- Douglas L. Smith, Editor, E&S magazine, pilation. I enjoy every issue of Caltech Public Relations, Mail Code 1-71, Pasa- E&S, but this one was special. dena, CA 91125, or e-mail [email protected]. We reserve the right to edit any letters selected for Walter V. Goeddel (MS ’51) publication for length, content, and clarity.

8 engineering & s c i e n c e n o . 1 39 40 engineering & s c i e n c e n o . 1 8 Obituaries

S e y m o u r Be n z e r 1921 – 2007

crystal for a secret military infected E. coli, their offspring project. His work led to the contained new genes that first transistor and a Nobel combined elements of the Prize for William Shockley same gene from both progeni- (BS ’32), John Bardeen, and tors. Benzer analyzed tens of Walter Brattain in 1956. The thousands of these so-called researchers who developed the recombination events, in transistor came to him and which portions of the gene told him, “You should have called rII swapped places. done this,” he recalled. He By comparing the length of said, “It escaped me, and it these portions, he mapped was under my nose.” rII’s fine structure, showing By then, however, he was an that it was not an indivisible assistant professor in phys- unit of heredity, but many ics at Purdue with a renewed smaller units strung together. interest in biology. Inspired His map was on a scale large by Erwin Schrödinger’s book enough, in fact, to see changes What is Life?, Benzer attended the size of a single nucleotide Seymour Benzer, the science from an early age. a summer course in 1948 —the letters that make up Boswell Professor of Neuro- During summer trips to the on bacteriophages, viruses the DNA code and formed science, Emeritus, a founder Catskill Mountains in upstate that attack bacteria, at Cold the double helix. This work of modern genetics, and one New York, he caught and Spring Harbor—organized bridged the gap between clas- of the giants of 20th-century dissected frogs. A whole by Caltech biology profes- sical genetics and molecular science, died from a stroke new world opened up when sor Max Delbrück. “Three biology. on November 30, 2007, in his brother-in-law bought weeks of that, and I was As molecular biology Pasadena. He was 86. him a microscope for his converted,” Benzer said. He exploded in popularity, Benzer In a series of elegant experi- 13th birthday, he said in his continued studying bacterio- went in a different direction. ments, Benzer made ground- oral history. He looked at phages during a year at Oak In the 1960s, partly inspired breaking discoveries about everything he could find, Ridge National Laboratory by the divergent personalities the structure of genes, finding and did experiments in his in Tennessee before joining of his daughters, he became that they were not indivisible basement laboratory—such as Delbrück’s lab at Caltech for interested in behavior and units of inheritance, as many making frog legs twitch with two years as a postdoc. He the “nature-versus-nurture” scientists had believed. He electric wires. Although never then went to the Pasteur debate. He began experi- also pioneered the field of a religious person, Benzer Institute in Paris for a year, ments with the fruit flyDro - behavioral genetics, in which said he respected his parents’ before returning to Purdue. sophila melanogaster while on he probed the connection faith, and followed his father Back in Indiana, he started sabbatical from 1965 to 1966 between genes and behavior. to synagogue on holy days. what would become some of in the lab of Roger Sperry, Benzer was born to Polish There, he would slip a physics his most well-known work on Caltech’s Hixon Professor of immigrants in New York City, book on top of the Torah; his the structure of the gene. Psychobiology, and stayed growing up on a street nestled father looked the other way At the time, molecular on to become a professor of between Jewish and Italian while he read. biology was in its embryonic biology in 1967. At Caltech, neighborhoods in Brooklyn. Benzer was the first in his stage. James Watson and Benzer developed a novel During the Depression, his family to go to college, enroll- Francis Crick had just discov- device that allowed him to parents, who worked in the ing in Brooklyn College in ered the double-helix struc- separate flies according to garment district, managed to 1938. Although he was inter- ture of DNA in 1953. But behavior and isolate mutant shield him and his three sisters ested in biology, he eschewed until Benzer’s experiments, strains. Benzer treated the from most of the era’s hard- the introductory classes and the physical nature of the gene flies as if they were particles, ships. To make ends meet, instead graduated with a was a mystery. bits of statistical data from Benzer’s parents brought physics degree in 1942. He He worked with mutant hundreds of individuals that clothes home to work on late went on to earn his PhD at strains of a bacteriophage he could collect in minutes, into the night. Purdue University, developing that infected Escherichia coli. rather than taking weeks to Benzer had an interest in a special type of germanium When two strains of the virus prepare a handful of rats.

8 engineering & s c i e n c e n o . 1 41 D a v i d C. El l i o t 1917 – 2007

His lab first studied the flies’ famous “Methuselah” gene in such as rotten fish or response to light, creat- fruit flies. Named after the chocolate-coated grubs. ing strains that failed to go Biblical character who sup- His enthusiasm for trying toward a light, as normal flies posedly lived 969 years, the diverse cuisine paralleled his do. Benzer, his students, and gene is key to longevity. The passion for reaching across his postdocs also developed mutants lived 35 percent lon- scientific disciplines, having strains that slept and woke ger, tolerated higher tempera- gone from physics to molecu- at random intervals, flies tures, survived longer without lar biology to neurogenetics. that died early, and mutant food, and were more resistant Colleagues remembered him females that brushed away to poison than normal flies. as a visionary and scientific males. By finding these kinds Over his career, Benzer maverick, following wher- of mutants, they identified accumulated more than 40 ever his curiosity took him. the genes responsible for the honors, including member- But his science was more flies’ circadian rhythms—the ship in the National Academy than mere interest—it was natural biological clocks of of Sciences, the Royal Society, an extension of who he was. organisms—and other genes and the American Academy Often waking up just before David C. Elliot, professor responsible for courtship, of Arts and Sciences. He won noon, he would work deep of history, emeritus, died on memory, and learning. the National Medal of Sci- into the early morning hours, November 21. He was 90. His research was contro- ence, the Wolf Prize in Medi- prompting colleagues to won- Born in Larkhall, Scot- versial at the time, as many cine from Israel, the Crafoord der if his own biological clock land, Elliot was the son of a scientists were skeptical as Prize of the Royal Swedish was the inspiration behind his minister. He attended the to whether the small and Academy of Sciences, the research on circadian rhythms. University of St. Andrews, simple fruit fly could be used International Prize for Biology He relished starting afresh in where in 1938 he met Nancy to dissect the complexities of from Japan, the Albert Lasker a new field in which he was Haskins, an exchange student behavior. His first seminar in Award for Basic Medical ignorant and could ask basic from Chattanooga, Tennessee. Sperry’s lab outlining some Research, and the Albany questions. “Ask stupid ques- The next year, they decided to of his initial fly research was Medical Center Prize. He also tions,” he said, “and you often get married. met with a divided reaction. won the Gairdner Interna- get amazing answers.” India was then under Brit- “They were pretty much split tional Award twice. In 2000 Benzer is survived by his ish colonial rule. Elliot joined down the middle between he became the subject of the wife, Carol Miller; two daugh- the Indian Civil Service, those who thought that this book Time, Love, Memory: A ters, Barbara Freidin and of which only about 1,000 was great stuff and others who Great Biologist and His Quest Martha Goldberg; a son, Alex- members governed the coun- thought this was pure crap,” for the Origins of Behavior, by ander Benzer; two stepsons, try at a given time. In 1940, he recalled. “They were really Jonathan Weiner. Although Renny and Douglas Feldman; he was sent to the Punjab screaming at each other.” many colleagues said his and four grandchildren. His region at what today is the Nevertheless, Benzer was work deserved a Nobel Prize, first wife, Dorothy Vlosky, northern border of India and highly respected, and he the award eluded him. “My died in 1978. —MW Pakistan. At 23, he governed pursued his interests with mother always regarded me an area larger than Scotland, freedom. His work with fruit as a failure because I didn’t and he spent much of his time flies grew into the new field of get the Nobel Prize,” he traveling the countryside on neurogenetics, showing that remarked. horseback. He would return much of behavior is hardwired Benzer savored gastronomi- to India in 1997 for the 50th and not the result of one’s cal experiences, whether it anniversary of the nation’s environment. was discovering sushi in Japan independence. He became Caltech’s or, upon encountering a In order for Haskins to join Boswell Professor in 1975 dearth of good restaurants in Elliot in India, she had to be and officially retired in 1992, Cambridge, England, forming a British citizen—or at least although he remained an a gourmet club with friends to married to one. She went to active researcher afterward. prepare their own meals. Col- South Carolina, which recog- In the late 1990s, Benzer leagues in his lab recalled him nized common law marriage, and colleagues identified the offering them bizarre food and married Elliot by proxy.

42 engineering & s c i e n c e n o . 1 8 Get credit for being part of the world's foremost scientific community with a CEFCU VISA card. Combining a choice of four stellar designs with solid, down-to-earth value, each card is a proud reflection of our membership. A CEFCU VISA offers: • A low rlX8drllUld 10.5% APR • NO annual fee • NO ca.h advance fee • 25·day grace period And our cards feature high credit limits and low minimum monthly payments. You've earned this VISA card. Use it with pride. Choose it to save. Make it yours today!" For more information, call us or visit a CEFCU office location.

800.592.3328 • _.emu.org

Foothill OffIce: 528 Foothill Blvd .• La Canada Flintridge JPL Office: 4800 Oak Grove Dr., Bldg. 218 • Pasadena Camp... Officez 515 S. WLison Ave .• Pasadena

• All endjt ...b;sct to ~8 tmJ ",.""bsnbipengineering d;gibiJj~ Rstu & tmJs c ..... i e n... c ebjKt n oto. cbtmgs 1 1IIitbottttuitbota43 MJria.MJric& CEFCU"CEFCU;' (1;11 EqIml Opportlmiry Letukr. She spent the next six years trying to join him in India, H e r b e r t B. Ke l l e r finally succeeding in 1945. 1925 – 2008 They had a second wedding in Lahore. The couple left India in 1947 and went to the United Keller made significant and double centuries, which States, where Elliot stud- contributions toward tech- are rides stretching 100 or ied history at Harvard. He niques for solving complex 200 miles; he said he rode his earned a master’s degree in problems with a computer. last double century when he 1948 and his PhD in 1951. He was known for developing was 72. He later earned a second methods to solve two-point In addition to serving on master’s from Oxford in 1956. boundary-value problems, numerous committees and Immediately after Harvard, he which arise in many areas of councils, he was a fellow ventured out west to Caltech, physics and engineering, from of the American Academy became an assistant profes- fluid flow to stellar structure. of Arts and Sciences, the sor, and was appointed a full He also made strides in bifur- American Association for the professor in 1960. He served cation theory, which looks at Advancement of Science, and as secretary of the faculty from how changes in parameter val- the Guggenheim Founda- 1973 to 1985, before retiring ues influence a system. One tion. He was the president of in 1986. simple example is the problem the Society of Industrial and Elliot’s research interests of how changing the number Applied Mathematics, and included the Liberal Party in of fishing licenses given out later won their von Kármán Scotland, the English Res- Herbert B. Keller, profes- each year affects fish popula- Prize. With Eugene Isaacson, toration, arms control, and sor of applied mathemat- tion dynamics. He remained he coauthored a textbook that national defense. A con- ics, emeritus, and a leader an active researcher even after became a classic in numerical sultant for RAND, NASA, in numerical analysis and his retirement in 2000. analysis. and the Ford Foundation, he scientific computing, died in Colleagues described him His brother, Joseph, a served as a trustee and honor- his Pasadena home on January as a mathematician with retired professor of mathemat- ary trustee of the Institute 26, after his routine morning chutzpah, unafraid to speak ics and mechanical engineer- of Current World Affairs, bicycle ride. He was 82. his mind and to go after ing at Stanford University; and he spent 30 years on the The son of a bartender who whatever problem interested his son, Steve; his daughter, board of trustees of Westridge loved numbers and puzzles, him—advice that he doled Debra; and four grandchil- School in Pasadena. He was Keller was born in Paterson, out through the years as an dren survive him. —MW chairman of Caltech’s 75th New Jersey. He studied elec- influential mentor to dozens anniversary celebration and, tronics at Georgia Tech and of students and postdocs. His in 1977, Caltech students joined the Naval Reserve Offi- fearless approach to research voted him the most popular cers Training Corps. During mirrored his other passion in professor. On his own time, World War II, he became a life: cycling. Elliot was an avid golfer and fire-control officer in charge His brother recalled a bridge player. of the guns on the USS Missis- cycling trip they took in the Predeceased by his son, sippi, where he trained future south of France in 1948, John, in 1991 and his wife president Jimmy Carter to be when they inadvertently in 1994, he is survived by his a gunnery officer. joined the Tour de France daughters, Nan Elliot Hale Keller later went to New after riding through roads and Enid Elliot, a son-in-law, York University and received lined with cheering specta- Richard Kool; and four grand- his PhD in mathematics in tors. Keller rediscovered the children and two great-grand- 1954. He eventually became sport in the early 1980s, and children. —MW a professor of applied math- despite suffering countless ematics at the Courant Insti- accidents—many with seri- tute of Mathematical Sciences ous injuries—never stopped at NYU. In 1965, he came to riding. In one of his most Caltech as a visiting professor oft-told stories, he said a and returned as a full profes- collision he had with a pile sor two years later, joining of lumber in Germany fixed the newly formed applied- his nearsightedness. Typically mathematics group. He later biking 100 to 150 miles a became the executive officer week, Keller didn’t allow age for applied mathematics and to slow him down—he fin- director of Caltech’s branch ished a 1,250-mile European of the Center for Research on tour when he was 68. He Parallel Computation. completed several centuries

44 engineering & s c i e n c e n o . 1 8 JUST BECAUSE WE HAVEN'T BRAGGED DOESN'T MEAN WE DON'T HAVE SOMETHING TO BRAG ABOUT.

70%70% OF OUR MUTUAL FUNOSFUNOS AND VARIABLEVARIABLE ANNUITY ACCOUNTS EXXCCEEDEDEEDED THEIR MORNINGSTARMOR NINGSTAR MEDIAN OOVERVER THE PAST 3 AND 5 YEARS~YEARS ~

8888%% OF OUR ACTIVELACTI VELY MANAGED EQUITY AND LlFECYCLE PORTFOLIOS EXXCCEEDEDEEDED THEIR BENCHMARKSBENCHMARKS DURINDURINGG A VOLATILE 2007 MARKET.'

Food for thought when you're looking for a full-servicefull-service financial partner that can deliver consistent performance. Ask your financial provider how your funds compare.

Call 11-800· 800·-TIAA-CREF or visit tiaatiaa-cref-cref.org.org/ investmentslnvestments to find out about our performanceperformance..

~ ______~ I~~~~ I L-______~~~ ~------~FINANCIAL SERVICES I~~~~ J FOR~------~~ THE GREATER GOOD·

'The M(}mlngstarMQm l ~gstar median represents the midpoint ofof an inde~i n de~ of comparahlecomparable fundS/accounts. 'For' For the 33'j'e3r·year period,period , 6262%,%, forfor the 55·year·year period, 61%%.. RelativeRelatM! periormanceperlormance OYerOYer other perilXlsperioos may vary.vary. AllAll TIAA·CREFTIAA·CREF Investmentinvestment vehiclesvehicles are subjects~bJ ect to market and ollleother risk factors.factor.;. TIAA·CREF IInnddil'idualividua l & InstirutionalInstiMional Services,liC.Sel"llices,llC, and Teachers Personallfll/eslorsPersonal lmestors Services,Sefvlces.lnc" Inc .. Membe/SMembelS flANNRA,RA. disllibutedisUibute securities Pfooucts.pfoducts. CC4077140771 © 2008 TeachersTeachelS InsuranceInS