Michael Triantafyllou, a professor at the department of ocean engineering at
the Massachusetts Institute of Technology, and graduate student James Czarnowski
have built a prototype vessel called “Proteus the penguin boat”. Proteus is
powered through the water by two “foils” that hang over the stern of the boat
and look like rudders. The foils propel the boat forwards by mimicking the
movements of the pectoral flippers of penguins and turtles.

Each foil is powered by its own motor and is capable of two types of motion.
One is a side-to-side flapping, which Czarnowski calls a “heave”, and the other
is a twisting motion. The movements are synchronised by an onboard computer to
maximise forward thrust. In a submarine, the system could also make the craft
dive or rise.

Triantafyllou and Czarnowski’s penguin-like system is one step on from the
working model of the tuna fish they built while studying the mechanics of fish
propulsion (Technology, 1 October 1994, p 22). “Robo-tuna” was propelled by an
undulating body and flipper. Although it was a success, the moving body was
difficult to adapt for boats. “It was while watching the penguins at the New
England Aquarium that I realised nature had already developed the necessary
system,” says Czarnowski.

In tests, the flipper-like foils have an efficiency of 87 per cent compared
with 70 per cent for a propeller, says Triantafyllou. According to 1992 US fuel
and shipping statistics, improving efficiency by just 10 per cent on only 3 per
cent of vessels in the American fleet would save 120 million litres of fuel a
year, worth $15 million.

As a foil flaps, it sweeps through a greater area of water than the blades of
a propeller. Consequently, more water is thrown backwards, giving a larger
forward thrust. And whereas a propeller forces the water to rotate, wasting
energy, with the foils there is no rotation so energy losses are smaller. The
flipper-like foils have similar hydrodynamics to fish, which are very efficient
swimmers.

Proteus is just under 4 metres long and about 50 centimetres wide. It is
powered by two car batteries and controlled by a modest PC. It flaps its foils
about 200 times a minute to generate a top speed of 2 metres per second. Scaled
up to a full-sized ship, this would translate to a top speed of 30 knots, with
the foils flapping once every two seconds. Triantafyllou believes that the
propulsion method could be easily developed for any size of ship. “From our
studies, the system can accommodate the forces needed to drive an oil tanker,”
he says.

The system could be particularly useful in helping military submarines avoid
detection. Propellers generate a “signature” wake which can be used to locate
them, whereas the wake from the flipper setup looks more like that from a large
marine animal.

As the next step towards constructing a full-scale vessel, Triantafyllou and
Czarnowski are planning to build a version three times as large as Proteus at a
local boat yard. Triantafyllou thinks that a full-scale system will take about
three years to develop.

Neil Gershenfeld at the Massachusetts Institute of Technology and Isaac
Chuang of the University of California at Santa Barbara outline their scheme in
this week’s Science. Meanwhile, David Cory, also at MIT, Amr Fahmy at
Harvard University and Timothy Havel at Harvard Medical School will publish more
results in a forthcoming issue of the Proceedings of the National Academy of
Sciences.

In a hypothetical quantum computer, the 0s and 1s of the bits, known as
qbits, are represented by quantum states of particles, for example the up and
down spin of protons. Like the gates in an ordinary computer, quantum gates
operate on these qbits to calculate results as an array of new states.

What makes quantum computing special is that the array of quantum states
exists in a superposition of all possible states until it is observed. In
effect, when a quantum gate operates on this superposition state it operates on
all of these inputs at the same time—a process called quantum parallelism.
So a computation that would take thousands of years on a conventional
supercomputer might take only a few seconds on a quantum computer.

The snag is that this superposition can collapse into a single state if it is
disturbed. So research up till now has concentrated on sophisticated ways of
preserving superposition by isolating and cooling individual ions.

In contrast, the latest idea looks at millions of nuclei at once using a
nuclear magnetic resonance (NMR) spectrometer—a standard analytical
instrument found in thousands of laboratories.

Atomic nuclei exist in quantum states of up spin or down spin. These two
states can be used as the 0 and 1 of a bit and can be flipped by bursts of radio
waves.

The chemical environment of each nucleus shifts its spin energy slightly. So
each nucleus within a suitable molecule—and the caffeine found in a cup of
tea or coffee is a current favourite—can represent a different bit. Each
nucleus will respond to a slightly different frequency of radio energy, so they
can be addressed individually.

There are typically 10²³ molecules in an NMR sample, and this offers a way
round the problem of the superposition state collapsing. Instead of looking at
one molecule once, the NMR looks at all the molecules statistically over a
period of time, which preserves superposition.

The world market for chanterelles is worth some ÂŁ1 billion a year. All
200 000 tonnes of Cantharellus cibarius are collected from the wild.
The chanterelle belongs to a group of fungi called ectomycorrhizas, which form a
sheath of filaments around the roots of a host plant. The fungus is thought to
obtain the sugars it needs to grow from an exchange of nutrients involving the
host plant, the fungus itself and probably other microorganisms, too.

In this week’s issue of Nature, Eric Danell, now at the University
of Agricultural Sciences in Uppsala, Sweden, and Francisco Camacho who
supervised his PhD project on chanterelle cultivation and works at Oregon State
University in Corvallis report that they have grown chanterelles on pine
seedlings. Danell inoculated the seedlings with chanterelle tissue in a sterile
medium, and transferred them to a greenhouse after 12 weeks. Danell and Camacho
harvested their first mushrooms 16 months later. It used to be thought that
chanterelle mushrooms would only form on mature trees. Danell says that the
fungi take longer to establish themselves in the wild because few spores are
produced and many are not viable.

Danell suggests that inoculated seedlings could be sold to growers, who would
have to wait up to three years for the mushrooms to develop in their
plantations. But some fungus experts argue that Danell’s technique will still
not allow true mass production. Roy Watling of the Royal Botanic Garden in
Edinburgh says that other mycorhizzal fungi could oust the chanterelles from the
seedlings.

Danell concedes that his fungi, which were of Swedish stock, might not find
English, French or Italian soils to their liking. But he believes that careful
management of the soil in plantations or the culture of locally adapted
chanterelles should solve the problem.

Watling argues that the golden prize will be a way of growing mushrooms
indoors on artificial beds without a host tree. “If I can get that I’m sitting
on a yacht in the Mediterranean,” he says. Watling believes such a breakthrough
could come after several more years of research into the relationship between
the fungus, the plant host and any other organisms involved. But Danell
disagrees: “It is science fiction to believe you can produce chanterelles
without a tree.”

Until now, that is. Steven Lamoreaux at the Los Alamos National Laboratory in New Mexico has used a torsion pendulum connected to two surfaces to measure the force and has come up with a value which agrees with Casimir’s prediction to within 5 per cent.

According to quantum theory, empty space is not truly empty but contains virtual photons and other virtual particles which pop into and out of existence. In the space between two plates erected close together in a vacuum, only those virtual photons whose wavelengths fit into the gap a whole number of times can exist. So there must be fewer virtual photons between the plates than outside them, and this imbalance produces a small force that pushes them together.

Lamoreaux found that the Casimir force varied with the separation between the two surfaces. When they were 0.75 micrometres apart it was around one billionth of a newton (Physical Review Letters, vol 78, p 5). Ed Hinds of the Sussex Centre for Optical and Atomic Physics at Sussex University is enthusiastic about this result: “The experiment is a beautiful demonstration of an effect we all believed in.”

Cryptographic systems send information in a coded form that needs a “key” to
unlock it. This key—usually a string of ones and zeroes—must be sent
to the receiver separately. But in conventional systems, there is no foolproof
way of discovering whether an eavesdropper has intercepted the key or not. This
means that such systems can never be entirely secure.

The new approach, called quantum cryptography, relies on quantum physics to
get around this problem. It was demonstrated by Paul Townsend of BT Laboratories
in Ipswich and reported in Nature (vol 385, p 47). It relies on the fact
that at a quantum level, the mere act of measuring certain properties of a
particle changes those properties.

Townsend’s system sends information down an optical cable as a stream of
light pulses. The phase of each pulse could be 0, 90, 180 or 270 degrees. 0 and
90 degrees represent a one, while 180 and 270 degrees represent zeroes.

An eavesdropper has a choice of two tests—one to detect a pulse with a
phase of 0 or 180 degrees and the other for 90 or 270 degrees. If they choose
the wrong test they will change the phase of the pulse and the value of the
resulting bit may be wrong.

The legitimate receiver must examine each pulse and will also get some of
them wrong, so they send a message to the original author telling them which
test they used for each bit. The original author then replies by telling them
which bits to discard as suspect.

If the signal was tampered with, however, some of the phases and bit values
will be shifted before the legitimate recipient ever sees them. So even after
discarding some of the bits as instructed, the data in their final message will
be different from that of the original, and this can be picked up using standard
error correction procedures.

Richard Hughes of the Los Alamos National Laboratory in California has been
developing a similar system. He hopes to have a prototype running in the summer.
“The system is 100 per cent secure as long as the fundamental rules of physics
are correct,” he says. “In three years, there could be a working, commercial
system. Since Townsend’s work it is no longer a physics question but a
commercial one.”