TRAFFIC in Bangalore is now bumper-to-bumper just about everywhere and gridlock is a certainty during rush hour. A journey from the centre to the hub of the IT industry on the city's outskirts that took only 20 minutes a few years ago can now take two hours. Corporate limos jostle with autorickshaws, trucks, taxis and even vegetable vendors with pushcarts plying the day's produce. The city is choking under the influx of companies, both foreign and Indian, eager to partake of its seemingly inexhaustible supply of cheap programmers.

And there's little respite in sight for Bangalore's creaking infrastructure. You'd think companies would be starting to have qualms about opening new offices in the city. Think again. Some of the biggest names in IT are heading towards Bangalore once more, and this time round it's not cheap labour they are looking for. They are hunting down the brightest, most inventive minds in India to populate a swathe of cutting-edge research facilities.

The work being done in these labs rivals any in the US and Europe. Ajay Gupta, director of Hewlett-Packard's research labs in Bangalore, says India is the place to be. "HP sees its India lab as being on an equal footing with our other research labs worldwide," he says.

Things have moved on a long way in the 20 years since US chip giant Texas Instruments opened an office in Bangalore to crank out software for testing and verifying TI's chip manufacturing processes. Take the $80 million multidisciplinary centre set up by General Electric to serve the US company's research needs. The centre is GE's first and largest R&D lab outside the US. It has 2300 employees, 60 per cent of whom have a master's or PhD degree in science. Anything is fair game here - from plastics to turbines to molecular modelling. "We are not here to serve the Indian market, but to serve [GE's] global research agenda," says Guillermo Wille, managing director of the company's lab in Bangalore and the only non-Indian on site. The products of research done at the labs are purely for export.

Using techniques such as numerical analysis and computational fluid dynamics, the GE researchers have significantly improved the efficiency of the company's wind turbines and its engine for Boeing's planned 7E7 airliner. The centre is renowned for its materials science division, which invented a resin co-polymer that has made possible self-destructing CDs and DVDs. The discs are sold in sealed pouches: break the seal and the polymer reacts with air, making the disc unreadable 48 hours later. More mundane but no less significant is a water-saving washing machine also invented there
.
Another high-profile firm to set up shop in Bangalore is Google, the California-based company whose name has become synonymous with internet searching. Krishna Bharat, co-founder of Google Labs India and inventor of Google News, is looking for top PhD graduates from Indian universities to augment the dozen or so researchers working in Google's largely empty two-storey office in Bangalore, which opened last year. Bharat's team will research ways to improve internet searching in Indian languages and work on voice interfaces and other alternatives to the keyboard and mouse. Bharat expects his centre will soon contribute to Google's global research effort. Other high-tech giants that have opened research labs in Bangalore include Cisco, Intel, Sun Microsystems and Motorola.

While many of these companies' developments are intended for application worldwide, Hewlett-Packard's approach is different. Its Bangalore research centre, opened in 2002, has the express purpose of applying local brains to local problems. "The poorer people in countries like India aren't served by existing technology, so we need to find new technologies for them," Gupta says. His team is exploring new ways of making the internet accessible to non-English-speakers, and they have invented a Hindi language keyboard to cater to a majority of the non-English-speaking Indians. The lab has also developed a cheap touch-sensitive pad-based system to write emails. The text is digitised as you write and sent as an attachment to a normal email. "It opens up the possibility for people who are intimidated by keyboards to communicate via email," says Gupta.

“The biggest names in IT are hunting down the most inventive minds to populate their research labs”

Microsoft joined the party this year, with a research centre in Bangalore also intended to address the needs of India and other Asian markets, such as developing Indian-language versions of its software.
Companies are choosing Bangalore for one main reason: the availability of good computer-science professionals. "We weren't able to hire enough good-quality engineers in Silicon Valley," Bharat says. The concentration of high-tech companies in the city is unparalleled almost anywhere in the world. At last count, Bangalore had more than 150,000 software engineers - approaching the kind of numbers only Silicon Valley can boast.

As well as being a hotbed of computing expertise, Bangalore has significant scientific talent, especially in physics and materials science. It is this that companies such as GE have come to Bangalore for, along with Indian researchers' mathematical skills in analysis and modelling. "They spend as much time in front of the computer as they do in the wet lab," Wille says
.
“At last count, Bangalore had more than 150,000 software engineers, the kind of numbers only Silicon Valley can boast”

Though Bangalore is the main focus for high-tech in India, it is not the only one. Other cities are vying for a piece of the research pie. IBM, for example, set up its labs on the campus of the Indian Institute of Technology in New Delhi in 1998, and now employs 100 researchers. Besides contributing to global projects such as the WebFountain web search engine, the centre invented a comprehensive voice-to-text recognition system that translates from Indian-accented English and Hindi into text in the respective language. "It's only on rare occasions like with the speech recognition system that our work has application only in India. Generally we try to build generic solutions that apply globally," says Ponani Gopalkrishnan, director of IBM's Delhi labs.

Though foreign multinationals have dominated the research agenda in India till now, a growing number of Indians who have worked abroad are returning home with cash, contacts and confidence to set up companies of their own. Mouli Raman, co-founder of Bangalore start-up OnMobile says, "For the first time, Indians who have been exposed to the world realise they can do something just as good. They believe they can be world-class."

India special: The next knowledge superpower

THE first sign that something was up came about eight years back. Stories began to appear in the international media suggesting that India was "stealing" jobs from wealthy nations - not industrial jobs, like those that had migrated to south-east Asia, but the white-collar jobs of well-educated people. Today we know that the trickle of jobs turned into a flood. India is now the back office of many banks, a magnet for labour-intensive, often tedious programming, and the customer services voice of everything from British Airways to Microsoft.

In reality, the changes in India have been more profound than this suggests. Over the past five years alone, more than 100 IT and science-based firms have located R&D labs in India. These are not drudge jobs: high-tech companies are coming to India to find innovators whose ideas will take the world by storm. Their recruits are young graduates, straight from India's universities and elite technology institutes, or expats who are streaming back because they see India as the place to be - better than Europe and the US. The knowledge revolution has begun.

The impact of the IT industry on the economy has been enormous. In 1999 it contributed 1.3 per cent of India's GDP. Last year that figure had grown to 3 per cent. And what's good for one science-based industry should be good for others. India has a thriving pharmaceutical industry which is restructuring itself to take on the world. And biotech is taking off. The attitude is growing that science cannot be an exclusively intellectual pursuit, but must be relevant economically and socially. The hope among some senior scientists and officials is that India can short-cut the established path of industrial development and move straight to a knowledge economy.

For the New Scientist reporters who have been in India for this special report, many features of the country stand out. First, its scale and diversity. With a population of more than a billion, the country presents some curious contrasts. It has the world's 11th largest economy, yet it is home to more than a quarter of the world's poorest people. It is the sixth largest emitter of carbon dioxide, yet hundreds of millions of its people have no steady electricity supply. It has more than 250 universities which catered last year for more than 3.2 million science students, yet 39 per cent of adult Indians cannot read or write.

These contrasts take tangible form on the outskirts of cities from Chennai to Delhi, Mumbai to Bangalore. Here, often next to poor areas, great gleaming towers of glass are growing in which knowledge workers do their thinking. These images of modernity are a far cry from stereotypical India - a place bedevilled alternately by drought and flood, of poor farmers and slum-dwellers. Yet both sets of images are real - and many others besides.

High-tech is not the sole preserve of the rich. Fishermen have begun using mobile phones to price their catch before they make port, and autorickshaw drivers carry a phone so that customers can call for a ride. Technology companies are extending internet connections to the remotest locations. Small, renewable electricity generators are appearing in villages, and the government is using home-grown space technology to improve literacy skills and education in far-flung areas.

These efforts are often piecemeal, and progress is slow. "Illiteracy today is reducing only at the rate of 1.3 per cent per annum," says R. A. Mashelkar, director-general of the government's Council of Scientific and Industrial Research. "At this rate, India will need 20 years to attain a literacy rate of 95 per cent." He is hopeful that technology can speed up this process.

“Why spend large sums on moon rockets and telescopes when India still struggles to feed millions of its citizens? Because without science poverty will never be beaten”

Science too has its role to play. Critics of India's investment priorities ask why the country spends large sums on moon rockets and giant telescopes while it is still struggling to find food and water for millions of its citizens? The answer is that without science, poverty will never be beaten. "You cannot be industrially and economically advanced unless you are technologically advanced, and you cannot be technologically advanced unless you are scientifically advanced," says C. N. R. Rao, the prime minister's science adviser.

Rise of the middle class
The knowledge revolution is already swelling the ranks of India's middle class - already estimated to number somewhere between 130 million and 286 million. And the gulf in spending power between the poor and the comfortably off has never been more apparent. Take cars. Sales are rising at more than 20 per cent a year. Before India opened up its economy in the early 1990s, only a few models were available, almost all home-built. Today, top-end imported cars have become real status symbols. Another consequence of the knowledge revolution is that the extreme wealth of a new breed of young, high-tech yuppies is challenging traditional gender roles and social values.

Whether the new-found prosperity and excitement of present-day India can be sustained will depend crucially on how the government guides the country over the next few years. Cheap labour and the widespread use of English do not guarantee success, and there are major obstacles that the country will need to tackle to ensure continued growth. Take infrastructure. Where China has pumped billions into water, road and rail projects, India has let them drift. Likewise, companies complain that bureaucracy and corruption make doing business far more difficult than it ought to be.

One of the critical issues facing India is the gulf between the academic world and industry. The notion that scientific ideas lead to technology and from there to wealth is not widespread. This stems in large measure from the attitudes prevalent before 1991. Before economic liberalisation, competition between Indian companies was tame, so they were under no pressure to come up with new ideas, nor did academics promote their ideas to industry.

India's attitude to patents are a product of that mindset. The country has no tradition of patenting, and only recently have institutions and academics started spinning off companies and filing for patents in earnest. Most applications filed in India still come from foreign companies. Until this year, the country did not recognise international patent rules, a failure that hampered interactions with foreign companies.

The suspicion remains that Indian companies are out to steal ideas, says Gita Sharma, chief scientific officer of Magene Life Sciences, a start-up company in Hyderabad. "We are not yet able to wipe away that image." And while India has now adopted those international rules on paper, there are still concerns about how strictly they will be enforced. "It will take a couple of years before the full implications play out," says Sankar Krishnan, a biotechnology analyst for McKinsey and Company in Mumbai.

Bringing research round to a more commercial way of thinking is not the only issue that academia must face up to. Another cultural problem, according to some scientists, is that too often institutions have an ethos of playing safe. Researchers who devise and test daring theories are criticised if they fail, discouraging the kind of ground-breaking research that India needs.

There is a widespread view that the entire university system needs an overhaul. India awards only 5000 science PhDs a year, says Mashelkar, yet it should be producing 25,000. There are funding problems and political interference in the running of some universities, particularly those run by state governments. In response, central government has decided to select 30 universities, give them extra money, and mentor and monitor them to create a series of elite institutions.

But such changes will be for nothing if students choose not to study science. In recent years, increasing numbers have chosen to study IT and management because that's where money is to be made. "IT and outsourcing has improved the economy and quality of life of people, but has had a negative effect on science," Rao says. Mashelkar hopes that as science-based companies grow, and demand for fresh blood increases, salaries will rise and more students will opt for science.

Chasing China
These problems must be solved if India is to capitalise on its recent gains, and there are hopeful signs that Indian science is improving in the global scheme of things. Its share of the top, highly cited publications has increased, but it is starting from a very low base. The government spends only $6 billion a year on research and it still has fewer scientists per head of population than China or South Korea.

India's greatest rival has always been its giant neighbour to the north. While IT and services are helping India log 6 per cent year-on-year increases in GDP, China's vast manufacturing base is raising its GDP by around 9 per cent a year. Even in India's strong suit of knowledge-based industries, China could still steal the march on it, not least because its Communist government can command change, while in India the democratic government can only guide national development.

Nevertheless, the rewards for India of a thriving science-based economy could be huge. The investment bank Goldman Sachs estimates that if India gets everything right it will have the third largest economy in the world by 2050, after China and the US. India is not yet a knowledge superpower. But it stands on the threshold.

India special: The IT pioneer
Interview with Mr Narayanamurthy, Chairman of Infosys, a billion dollar plus Indian IT firm

What was world's perception of Indian programmers when you started?
There was no perception. It was an uphill struggle to convince foreign clients that we could do what they wanted.

How did you do that?
At the heart of our pitch were India's strengths as a nation. The widespread use of English, the large technically trained workforce and what seems to be an innate talent for computer programming all helped. For Infosys the social changes that were sweeping India, such as a new-found drive for entrepreneurship and the liberalisation of domestic markets, also provided a shot in the arm.

What makes Indian software development stand out?
The Indian way of working. This means breaking down big software projects into chunks that can be tackled simultaneously by many groups worldwide, and then drawing them back together to create a single product. We needed new management processes and new software tools - Infosys pioneered this way of working.

How has Infosys helped change India's image overseas?
We have gone from an era when it would have been inconceivable for Indian companies to compete with western companies, to a time when we are fast becoming one of the most important technology development centres in the world, all in the space of two decades.

Many western companies have opened operations in India. Has this changed the sort of work done here?
Yes. Large western companies, such as Texas Instruments and Intel, are doing very sophisticated work in India - they are building the next-generation chips. Many venture capitalists in Silicon Valley now insist that development be done in this country. That means a significant section of the next generation of technology companies will have research facilities here.

What will the consequences of this change be?
We are creating an alternative innovation centre to those in the west because there is intellectual muscle, money and a growing domestic market here. That should help stem the brain drain to the west.

There will also be an impact on students overseas. For example, US students will worry that IT jobs will migrate to India and so will stop studying technical subjects. They are already becoming wary of going into a field which will be "Bangalored" tomorrow.

If the immigration of technology experts to the US and the supply of their graduates eventually dry up, then I can imagine the centre of gravity of innovation drifting east.

What could derail this process?
Indian infrastructure. For example, outside our campus, Bangalore is straining to keep up with the pace of change the IT industry is forcing on it. The environment has not developed the ability to respond to the change.

India's problems are diverse and daunting. How can IT help?
None of us claims that the IT industry is a panacea for India's problems, it would be ludicrous to say so. But what we have done is significant. It is a necessary condition for progress, because our actions have had an impact globally and have shown the power of human capital in India. The growth of the industry has shown us what can be: it has shown us the change that we need in this society. We need to use the momentum to change other parts of our society. That's the leap in leadership which is missing.

India special: Space programme presses ahead

Rocketing evolution
NESTLED amid the eucalyptus, cashew and coconut trees of Sriharikota Island on the eastern coast of India, north of Chennai, is a 76-metre steel tower. If all goes to plan, some time in late 2007 the tower will be engulfed in flames as India's first mission to the moon blasts off. Sriharikota will also be the launch site for India's most advanced scientific research satellite, Astrosat. The satellite will measure, among other things, X-ray radiation emitted by matter sucked into black holes and given off at the birth and collision of stars.

But why is India, a country that still has so many development problems on the ground, aiming for the heavens? To Indian scientists, the question is not only patronising of their scientific aspirations, it betrays an ignorance of the Indian space programme's greater purpose and successes against the odds.
India's political leaders say the country cannot afford not to have a space programme. Indira Gandhi, who was India's longest-serving prime minister, believed it was not only important for science, but also vital to India's development.

“Several groups are trying to create artificial life by identifying the essential genes needed to support a simple living cell, and then stitching together this minimal genome”

Take, for example, India's six remote-sensing satellites - the largest such constellation in the world. These monitor the country's land and coastal waters so that scientists can advise rural communities on the location of aquifers and where to find watercourses, suggest to fishermen when to set sail for the best catch, and warn coastal communities of imminent storms (see "Eyes in the sky"). India's seven communication satellites, the biggest civilian system in the Asia-Pacific region, now reach some of the remotest corners of the country, providing television coverage to 90 per cent of the population. The system is also being used to extend remote healthcare services and education to the rural poor.

But it has been a long time coming. When India first detonated a nuclear device in 1974, the US and European nations imposed widespread sanctions to restrict India's access to technologies that could be used to make a nuclear missile. This hobbled the country's rocket development programme and forced the Indian Space Research Organisation (ISRO) to reinvent technologies it could no longer buy. In the long run this has given India an advantage over other countries with aspirations to reach space. Its space programme is already largely self-sufficient and aims to soon be completely independent of foreign support.

It hasn't all been plain sailing. Its first rocket, the Satellite Launch Vehicle, ended up crashing into the Bay of Bengal 5 minutes after launch in 1979. The following year it placed a 40-kilogram remote-sensing satellite into near-Earth orbit, but the satellite's perigee was lower than planned and it entered the atmosphere and burnt up after only 130 orbits. The Augmented Satellite Launch Vehicle followed, and in 1992, after two crashes, it finally succeeded in lifting 150 kilograms to a height of 435 kilometres.

“India's first rocket, the Satellite Launch Vehicle, ended up crashing into the Bay of Bengal 5 minutes after launch”

ISRO moved on to the Polar Synchronous Launch Vehicle, designed to carry 1-tonne satellites to a height of nearly 1000 kilometres. Though the first launch in September 1993 crashed, the PSLV has since performed flawlessly, placing seven Indian satellites and four from other countries into orbit. "It is our workhorse," says Madhavan Nair, director of ISRO. In 2003 the rocket was used to launch India's latest remote-sensing satellite, IRS-P6, capable of imaging the Earth's surface to a resolution of 5 metres.

To put its heavier communications satellites into geostationary orbit, India still has to rely on foreign hardware. But that may soon change. ISRO's latest rocket, the Geosynchronous Launch Vehicle (GSLV) is able to lift large satellites into geostationary orbit, 36,000 kilometres up. On 20 September 2004, the GSLV launched the 2-tonne EDUSAT, the world's first satellite dedicated to providing support for educational projects.

One of the GSLV's rocket boosters is a Russian-made cryogenic engine. International sanctions meant India was only allowed to buy engines, not the know-how to design and build them. So for future rockets ISRO engineers are developing their own. Ground tests have been completed and the plan is to launch a completely home-made GSLV-Mark 2 by the end of this year, Nair says.

ISRO is already planning the next-generation GSLV, the Mark 3, which will be powerful enough to launch India's biggest satellites. Nair now has his sights on the commercial market. A launch on GSLV-Mark 3 should cost about half the rate charged by France, the US and Russia, he says.

“The moon probe will map the lunar surface at resolutions down to 5 metres, something that has never been done before”

India's space programme is already a money-earner. ISRO sells infrared images from its remote-sensing satellites to other countries, including the US, where they are used for mapping. And the Technology Experiment Satellite, launched in October 2001, is beaming back images of the Earth's surface with a resolution of 1 metre, though they are not yet available commercially.

Three per cent of ISRO's $3.3 billion 5-year budget is devoted to the planned moon mission. A reconfigured PSLV rocket will lift Chandrayan - "moon vehicle" in Hindi - to 36,000 kilometres, after which the craft's own engines will take it to the moon. Nair says one of the purposes of the mission is to inspire Indian youngsters to take up a career in science.

Chandrayan will create 3D maps of the moon's surface at a resolution of between 5 and 10 metres, something that has never been done before. It will also map the distribution of ilmenite, a mineral that traps helium-3, a possible source of energy for future bases on the moon. No manned missions are planned, but if the trip is successful, robots might be sent up to collect samples.

According to Nair, the Madras School of Economics in Chennai has estimated that ISRO's projects have added between two and three times the organisation's budget to the nation's GDP. Several countries in Africa and Asia are seeking ISRO's help to emulate the model. "India is perhaps the only country where societal needs are met by the space programme in a cost-effective manner and the services are reaching the needy," says Nair.

Eyes in the skyTHOUGH Astrosat and the moon mission are the headline-grabbers, using imaging satellites for development remains at the top of the agenda for the Indian Space Research Organisation (ISRO).

For instance, when the skies are relatively clear between January and March, infrared images are used to measure the reflectance of plant-covered surfaces to check how well watered the crops are. It is also possible to distinguish between crops such as rice, wheat and cotton, and even predict whether a crop might fail for lack of water. "We are able to forecast the yield one month before the harvest," says Venkatakrishna Jayaraman, the director of ISRO's Earth Observation System. In this way the government can be forewarned of possible food shortages.

Ensuring a supply of clean drinking water is a problem in many parts of rural India. Villagers often resort to guessing the right spot to drill a well based on experience, but it is a hit-and-miss affair. Topographic and hydrological maps produced from satellite images allow Jayaraman's team to help rural communities locate areas most likely to yield underground water. "The success rate for drilling wells has gone up from 45 to 90 per cent," says Jayaraman.

The next step for Jayaraman's team is to use the same information to work out where to build small dams to capture rainwater and recharge underground reservoirs. This approach could help reclaim arid and semi-arid land for agricultural use.

ISRO's satellites are also having an impact at sea. OCEANSAT, launched in 1999, monitors the chlorophyll content of oceans and the sea surface temperature. ISRO scientists use the information to identify areas where cold, nutrient-rich water wells up from the ocean floor, which in turn attract fish.The coordinates of these areas are then sent to more than 200 coastal centres. The upwellings last for several days - meaning the areas identified contain high concentrations of fish - long enough for fishermen to arrive and gather a sizeable catch. According to Jayaraman, fish catches have doubled in the last decade.

Besides remote sensing, ISRO operates eight communications satellites. These are now used by 35,000 commercial customers, all based in India. "If we didn't have the satellites, they would have gone and hired a satellite from somewhere else," says Appana Bhaskarnarayana, director of ISRO's satellite communications programme.

ISRO has also used these satellites to implement disaster-warning systems. In 1977 a cyclone killed 10,000 people on the coast of Andhra Pradesh in south-east India. In the 1990s data from meteorological satellites was used to warn of a similarly devastating cyclone, dramatically reducing the loss of life to 900. "The technology has helped to predict cyclones maybe a few hours in advance," says Bhaskarnarayana. "Lots of lives have been saved."

ISRO is becoming more ambitious in how it plans to use these satellites. It has already linked 69 hospitals in remote areas of India such as the Andaman Islands to 19 hospitals in India's main cities. A health worker in a remote location can then transmit a patient's medical information to a specialist in seconds and, in many cases, a video consultation is sufficient for diagnosis. This means the patient can avoid travelling huge distances unless it is absolutely necessary.

India special: Vaccines for pennies

WHEN Krishna Ella went to venture capitalists in 1995 he was laughed out of their offices. A molecular biologist at the University of Wisconsin in Madison, he was proposing to make hepatitis B vaccine in India, his native land, for a mere dollar a shot. At the time UK drug firm SmithKline Beecham was selling the product in the west for 20 times that amount. "People thought: there's no way this guy can produce this vaccine at a dollar," Ella recalls.

Undeterred, he and his marketing manager wife Suchitra Ella sold their houses in America and India, abandoned their US careers and left for Hyderabad to set up their own company. They sank all they owned into the venture, begged money from friends, and finally won backing from an Indian bank. Their company, called Bharat Biotech, now sells the vaccine in developing countries for 28 cents a shot. It owns the second biggest production facility for this vaccine in the world and has an annual turnover of $7.3 million. "Those venture capitalists are kicking themselves now - you bet!" says Suchitra Ella.

The pair are typical of highly educated Indian expats who have forged their careers in the west, but are now returning to take advantage of new economic opportunities at home. Their intimate knowledge of western science and business is invaluable, and they are natural risk-takers. After all, this is not the first time in their lives they have made a daunting fresh start. "There is always dogma in science," says Krishna Ella. "To break the dogma you need to take risks."

Krishna and Suchitra Ella wanted to give something back to their home country, and setting up in India made good financial sense too. The country's economic liberalisation in the early 1990s has led to a wealth of new business opportunities and there is also a ready supply of well-educated scientists who are less costly than their US counterparts. And some states are throwing money at biotech start-ups. Genome valleys and knowledge parks are the current vogue as politicians try to tempt entrepreneurs with tax breaks, simplified regulation and guaranteed high-quality water, power and communications links.

So how did the Ellas manage to undercut the competition so dramatically? Hepatitis B is caused by a virus that attacks the liver, and which can cause lifelong infection, liver failure and cancer. It is usually spread through sex with someone who is infected, or by drug users sharing needles. In 1986 SmithKline Beecham launched a vaccine for hepatitis B, the first in the world to be produced by genetic engineering. It is made by adding genetic material to yeast cells so that they produce a key protein from the surface of the virus. People who are immunised with that protein produce antibodies that protect them should they subsequently encounter the virus.

Ten years ago Krishna Ella spotted that the purification method SmithKline Beecham used to extract the vaccine protein was relatively inefficient and costly. The multinational was using - and still uses to this day - a technique called ultracentrifugation, in which samples are subjected to 100,000 times gravity to separate the protein from DNA.

“Krishna Ella spotted that the method being used to extract the vaccine protein was inefficient and costly”

The equipment cost over $1.5 million and only recovered 15 per cent of the protein. What is more, the technique used caesium chloride, which is expensive and has to be completely removed from the final product because it is toxic. That makes disposal costly too.

Krishna Ella had come up with a new purification process that would eliminate ultracentrifuges and caesium chloride, and boost efficiency to 80 per cent. The vaccine protein has a phospholipid tail that is electrically neutral, unlike most of the yeast proteins and DNA, which carry an electrical charge. With Ella's method, called the Himax technique, the vaccine protein can be made to precipitate out of the solution onto a special matrix, while all the charged molecules stay put.

Bharat has since started manufacturing other products, such as a typhoid vaccine and an antibiotic for use against staphylococcus bacteria, which can cause skin and blood infections and pneumonia.

And the firm has got funding from the Bill and Melinda Gates Foundation to carry out malaria vaccine research and to develop a cheap vaccine for rotavirus, a major cause of childhood diarrhoea in poor countries. This type of diarrhoea kills about half a million children every year.

Krishna Ella says he wants to tackle third-world diseases neglected by the multinationals, a sentiment often voiced by Indian entrepreneurs who believe scientists have a duty to the poor.

"It feels very satisfying," says Suchitra Ella. "We are on top of the world because we are doing something that is really required for countries like India."

India special: Radio telescope offers dishes to savour

"I'm in love with this," says Subramaniam Ananthakrishnan, sweeping his arm over a rural landscape dominated by giant dishes. The radio astronomer has reason to be besotted. He and his colleagues spent 15 years transforming this once desolate region of Khodad, 90 kilometres from Pune in western India, into a home for the Giant Metrewave Radio Telescope (GMRT) - the world's largest, low-frequency radio telescope and India's biggest basic science project.

This is Big Science on anyone's scale. The telescope consists of 30 antennas, each one 45 metres across. Twelve of them sit in a 1-kilometre-square central region, while the remaining 18 stretch out along three arms, each 14 kilometres long. The central cluster allows the telescope to pick up extremely faint signals, and the arms give it high resolution.

The GMRT opens a unique window. Its operating frequencies - between 130 and 1500 megahertz - are at the opposite end of the spectrum from gamma-ray astronomy. "These frequencies have been looked at before, but either the resolution or the sensitivity was not there," says Rajaram Nityananda, director of the National Centre for Radio Astrophysics (NCRA) in Pune. "That combination is unique to the GMRT."

India is a good spot for this world-class facility. The country has a long tradition of radio astronomy and, compared with many rich nations, its airwaves are relatively uncluttered. Through clever innovation, such as using a mesh of fine wires to form the reflecting surface of each dish, Ananthakrishnan and his colleagues, led by NCRA's Govind Swarup, have created a revolutionary, low-cost design. The entire telescope cost $12 million.

Since the telescope was completed in June 2000, astronomers from around the world have been jostling for time on GMRT. But this happy outcome was not guaranteed, remembers Ananthakrishnan. In 1985, while searching for a site, the astronomers surveyed a region surrounded by lush sugar cane fields in Dhond, about 100 kilometres from Pune. Suddenly, their jeep was surrounded by villagers wielding knives and axes. "They told us that if we tried to purchase their lands via the government, they would quickly dispatch us to heaven," he says. "We saw their point and left Dhond in peace." Later, they settled on Khodad, a once-barren region now fast becoming India's wine-growing centre.

One of the GMRT's main tasks is investigating the clouds of hydrogen gas thought to be the precursors of galaxies. Its biggest asset is its ability to detect a 1420-megahertz radio signal emitted by excited hydrogen gas. In distant galaxies, which are moving swiftly away from us because the universe is expanding, this "spectral line" is shifted to a lower frequency by the Doppler effect. The shifted frequency is well within the telescope's range, allowing astronomers to use it as a probe for studying the dynamics of evolving galaxies.

NCRA astrophysicist Jayaram Chengalur used this technique to look at dwarf irregular galaxies (DIGs) which lack any apparent structure and are thought to give rise to larger galaxies. "Most people believed that very small dwarf galaxies just don't rotate at all," says Chengalur. His team measured the velocity of hydrogen gas in 51 DIGs with great precision (www.arxiv.org/astro-ph/0411664). "We found that very small galaxies do rotate," he says.

GMRT is also valuable for studying pulsars - dense, fast-spinning neutron stars that emit radio pulses at regular intervals. "GMRT can pick up pulsars that others who traditionally search at higher frequencies cannot pick up very easily," says NCRA astrophysicist Yashwant Gupta. Last year Gupta and his colleagues found a "binary system" comprising a pulsar circling another massive object. It has the most eccentric orbit ever seen (www.arxiv.org/astro-ph/0411664). The researchers' theory is that the pulsar first accumulated matter from a low-mass star, then somehow exchanged that for the more massive one it now orbits. Astronomers believe that a similar process could give rise to a pulsar-black hole binary system. "If you find a pulsar going around a black hole, that will be a fantastic discovery, because you can do all kinds of science with it," says Gupta.

Today, radio astronomers around the world are planning the next-generation radio telescope, which they are calling the Square Kilometre Array. "Everyone is looking at GMRTas a test area for the SKA," says NCRA astrophysicist Pramesh Rao. Radio astronomer Paulo Freire of Cornell University in Ithaca, New York, agrees: "The beauty of GMRT's design is deeply influencing the construction of the SKA."

India special: Embracing GM crops

"WESTERN protesters holding a cup of Starbucks have no business protesting against GM," says Kiran Sharma. Rich Europeans can afford to reject the technology, he says, "here, we don't have a choice."

Sharma believes passionately that GM crops can go a long way towards tackling hunger in the developing world. But he is no Monsanto stooge. Sharma is a scientist at the International Crops Research Institute for the Semi-Arid Tropics in Hyderabad, southern India. ICRISAT is a network of non-profit research institutes in developing countries, funded by donations from rich nations and international agencies.

GM succeeds where conventional breeding cannot, says Sharma, because it can produce traits, such as disease resistance and drought tolerance, that do not exist in a crop or its wild relatives. Bringing in genes from other species is the only way to improve these crops. "We are trying to give breeders something they don't have," he says.

India embraced GM in March 2002 when the government's Genetic Engineering Approval Committee gave the green light for three varieties of Bt cotton. The crops, owned by a Monsanto subsidiary called the Maharashtra Hybrid Seed Company (MAHYCO), have an added bacterial gene for a toxin that kills a major caterpillar pest called the American bollworm (Helicoverpa armigera). So far, Bt cotton is the only GM crop grown commercially in India.

“Kiran Sharma says GM technology allows him to beat diseases that traditional breeding has failed to tackle”

Advocates of Bt cotton say it lets farmers use less pesticide - typically one or two sprays per harvest as opposed to three or four sprays for conventional varieties. They argue this makes it cheaper and more environmentally friendly because the Bt toxin only kills moth and butterfly caterpillars. But no one has studied in detail the effect of the crops on non-target insects and other species.

MAHYCO claims the GM crop typically yields around 30 per cent more than non-GM crops, but critics dispute this. Suman Sahai is organiser of the anti-GM group Gene Campaign in New Delhi. She and colleagues studied 100 farming families growing GM and non-GM cotton in the states of Maharashtra and Andhra Pradesh. According to Sahai, yields of the non-Bt variety actually beat the GM crop by around 16 per cent, although the published results do not offer any figures to back up this claim.

Certainly that finding doesn't tally with the crop's popularity. "Farmers have bought it left and right," says Govindarajan Padmanaban, a biotechnologist at the Indian Institute of Science in Bangalore. "Farmers are cleverer than the activists or the companies. They won't buy things if they do not work."

Sahai's main objection is that embracing GM will hand over control of India's food supply to multinational companies that are motivated by profit rather than the best interests of farmers and consumers. "They have nothing in the pipeline that is targeting the poor," she says. "The public is completely excluded from the decision-making process." Why gamble on a potentially dangerous technology with economic risks, she asks, when old-fashioned selective breeding has served so well.

Sharma says GM technology allows him to beat diseases that traditional breeding has failed to tackle, such as clump virus and rosette virus, which infect groundnut plants. He is also working on a "golden" groundnut variety which manufactures extra vitamin A for a more nutritious crop. Sharma is now conducting small-scale field trials of GM groundnut, pigeon pea and chickpea engineered at ICRISAT (see "Staple crops go GM").

The chickpea and pigeon pea are both genetically engineered to contain a Bt toxin gene. Sharma began by producing lots of GM varieties differing from one another in the position of the inserted gene in the genome. This can affect how strongly the gene is expressed and how well it is transmitted to the next generation. Then he narrowed down the initial versions to the handful he is field-testing.

The aim of his present field trials is to discover which versions work best outdoors before moving on to large-scale trials in farmers' fields. Both chickpea and pigeon pea are naturally drought resistant and are widely grown for food by subsistence farmers. Ultimately, Sharma intends to distribute the GM seeds to farmers for free.

GM research only takes up around 10 per cent of the research at ICRISAT, but the researchers there feel they have a special contribution to make because they cannot be seen as being in the pocket of industry. "We see ourselves as the acceptable face of GM," says ICRISAT's deputy director-general, Dyno Keatinge.

There is an expectation among researchers that opposition to GM crops will melt away once their home-grown research begins to deliver tangible results. India's farmers are already voting for Bt cotton by buying the seed. GM crops that are "Made in India" can only get more popular.

Staple crops go GM

ICRISAT's palatial campus is an oasis of serenity after the noisy streets of Hyderabad. As Kiran Sharma drives me through part of the 1400-hectare site we pass fields of diminutive chickpea and pigeon pea plants next to imposing stands of pearl millet and sorghum. This haven, a half-hour drive from central Hyderabad, is home to 278 wild bird species, as well as monkeys and, slightly alarmingly, cobras. But I am here to see something that could change Indian agriculture.

Sharma stops the car next to a low fence. Within the small enclosure are rows of unimpressive-looking, knee-high plants. And in a central inner sanctum of netting designed to keep insects out are the world's first field tests of varieties of pigeon pea (Cajanus cajan). They have been genetically modified with the Bt gene, Sharma announces.

In an enclosure next door is a patch of bare earth, where Sharma tells me he planted another world first only the day before, Bt chickpea (Cicer arietinum). Both plants are grown primarily by poor subsistence farmers, but the conventional varieties are vulnerable to the American bollworm (Helicoverpa armigera), a caterpillar that can wipe out more than half a farmer's harvest. "These products are badly needed by subsistence farmers," says Sharma.

The non-GM plants in the outer enclosure act as a pollen trap: a way to find out if they pick up the inserted gene from plants in the inner sanctum and pass it to their offspring. They and the earth around them could be contaminated with GM pollen, so I am not allowed near them in case I then contaminate conventional varieties growing nearby.

Sharma's most advanced GM crop is a variety of groundnut (Arachis hypogaea) that is resistant to peanut clump virus, which can reduce harvests by 70 per cent. His team has inserted a gene for part of the virus's protein coat. The plants express the protein but do not fold it correctly, and for reasons Sharma is not yet sure of, this defective protein stops the virus from assembling its coat and escaping to infect other cells.

Groundnut is a particularly good candidate for genetic modification because it is almost entirely self-fertilised, so there is little chance of the foreign genes escaping. What's more, growing GM groundnut should benefit conventional growers in the area because the plant mops up virus particles in the soil. "Our transgenic plants are eliminating the virus," says Sharma.

James Randerson

Taking E-votes to the masses22 May 2004

The US could learn a thing or two from India's all-electronic national elections, which concluded last week. In the world's largest democracy, 650 million people had the chance to cast their votes on nearly a million e-voting machines. In contrast, only 20 per cent of the ballots to be cast in this year's US presidential elections will be electronic.

Controversy continues to dog the US's sophisticated e-voting machines. The complexity of the software makes it hard to spot bugs or any attempted fraud (New Scientist, 14 February, p 6).

However, a mere glance at the short, simple software in the Indian machines would reveal any bugs, according to Swami Manohar, a computer scientist at the Indian Institute of Science in Bangalore. The machines avoid features such as touch-screen balloting and transmitting votes to a central server. Instead, they merely record the number of times a button for a particular candidate is pressed. "I believe it was a terrific success," says Manohar.

Avi Rubin of Johns Hopkins University in Baltimore, Maryland, an outspoken critic of e-voting in the US, says, "I do believe that [the Indian] system is much better than the ones in the US, due to its simplicity." However, Rubin says the machines lack a paper printout of each vote that has been cast, which would allow voters to verify their choice of candidate, or scrutineers to do a recount if necessary.