Ageism Revisited!

Readers of the very first FUSION newsletter back in summer 2001 may recall an article about the age restriction of 27 imposed by the CERN Summer Student Programme. It ended with the words "We contacted them to ask why, but have not yet had a reply". So, did we ever get one, I hear you ask?

Well, after five months and several reminders, a reply did indeed appear. The Summer Student Programme, it said, was "designed to attract young people into the field of particle physics". I wrote again to ask whether they didn't perhaps mean "new" rather than "young" ... but at that point our already halting dialogue came to a complete stop.

Several months later, annoyed by the total silence, I decided to take up the matter with my MP. He passed my query on to Lord Sainsbury, Permanent Under-Secretary of State for Science and Innovation. As if by magic, the Minister managed to squeeze a response out of the faceless bureaucrats. (Perhaps he prefaced his letter with the words "About those 25 million euros we were going to give you..."?)

The main reason for the age restriction, Sainsbury was told, was that the scheme was "specifically designed for young students without previous work experience", and that it falls into the category of "programmes facilitating a student's integration into professional life", and is thus allowed to have an age limit.

So there you have it. You may have thought that CERN was a centre of scientific excellence, with physicists from all over Europe continually pushing back the frontiers of knowledge. Wrong. It is actually brim-full of young students doing management training workshops, learning how to write on flipcharts and making up acronyms, with perhaps the odd bit of first-aid and computer training thrown in, prior to taking up careers in finance and banking.

Personally I am grateful to Lord Sainsbury for clearing this little matter up. Those of us who have spent our working lives mending telephones, driving buses or bringing up children obviously already possess all the skills that CERN could have taught us. There is no need for the tedium of working alongside leading particle physicists engaged in cutting-edge research, to say nothing of having to jet across Europe and spend several months in the boring old Alps!

Jim Grozier

QUANTA AND CONTINUUM

Book Review

Nucleus - A trip into the heart of matter

By Ray Mackintosh, Jim Al-Khalili, Björn Jonson and Teresa Peña.

When writing anything it is important to know who your audience will be. At the FUSION AGM one of the authors of Nucleus (Ray Mackintosh) explained that this book was aimed at students in the upper years of secondary schools - intelligent sixth formers (ISFs). What they ended up with is a book aimed at ISFs who also own a coffee table.

Popular science books seem to fall into two classes. First there are those books where the emphasis is on the words - books by authors like Richard Dawkins, Ian Stewart - even Stephen Hawking's A Brief History of Time was (in)famous for the words therein. The other sort of popular science is the large hardback book, often writen as a tie-in to a TV series, where the stunning pictures seem to be the main attraction.

Nucleus gives the impression of starting out as the first type and then having pretty pictures added as a sudden conversion to the second type. Indeed the book is packed with interesting pictures. However, these pictures are rarely referred to in the text, and often seem to have only a tenuous connection with the topic being written about.

It even gives the impression that some useful (but less attractive) diagrams have been removed. When discussing nuclear energy levels we are told that they "...are often represented symbolically in diagrams by a series of horizontal lines." But there are no such diagrams in Nucleus. Maybe the book's art director should have actually read the copy.

So, what about the text? Well, the ISF had better have studied the subject first. It might be the result of having been written by a committee of four, but the topics in the first half of the book seem to have got jumbled during the publishing. Nuclear half-life is given three paragraphs in the chapter "Particles or Waves" and stands out like a sore thumb. Why it's not in with radioactivity is a sub-editor's mystery.

However, it does get better. The second half of the book is much more coherent. The variety and abundance of nuclei, applications of nuclear physics and how nuclear physics is responsible for much of modern astrophysics and cosmology are all covered in a way that will interest and educate the ISF. And the pictures are relevant to the text (although still not referenced therein).

As well as ISFs, this book would be excellent for those who have studied S103 and found the later physics blocks interesting and want to know more. It does not teach the fundamentals of nuclear physics, but higher level physics students will find it an interesting and worthwhile purchase - especially if they've got a space on their coffee table!

Frank Hollis

John Horlock Award

Each year the Association of Open University Graduates (AOUG) selects one of the Science Faculty's final or penultimate year PhD students, who must also be an OU graduate, to receive the John Horlock Award. The Award is named in honour of the OU's second vice-chancellor, and consists of a sum of money together with a commemorative scroll. Bob Lambourne tells us that AOUG has selected John Barker to receive the award for 2002. Congratulations to John and best wishes from FUSION.

Ten Million Years in a Photon's Life

"For the sun it actually takes a photon 10,000,000 years to get from the interior to the surface" -

Erika Böhm-Vitense, Introduction to Stellar Astrophysics. But is it the same photon? Hmmm, let's see.

Year 1, January 1st: 00:00 hrs.

I am born, as a result of a nuclear fusion reaction in the heart of the Sun. Hooray!

00:00 hrs 0.00000000001 sec.

Absorbed by an electron. Damn!

00:00 hrs 0.00000001 sec.

I am born, as a result of emission by an electron in the heart of the Sun. Hooray!

00:00 hrs 0.00000001001 sec.

Absorbed by an electron. Damn!

00:00 hrs 0.00000002 sec.

I am born, as a result of emission by an electron in the heart of the Sun. Hooray!

00:00 hrs 0.00000002001 sec.

Absorbed by an electron. Damn!

(To be continued)

S381 Astronomy Weekend

FUSION is planning a Weekend Practical Event for students of S381 - The Energetic Universe at the University of London Observatory in Mill Hill, north London (part of University College). It will take place on Saturday, March 22nd 2003 from 1.00pm to 11.00pm, and will include an opportunity to do some practical astrophysics, using the Observatory's five telescopes, if the sky is clear; if it isn't, and during the first half of the session, there will be activities based on "raw" data obtained on previous nights. The activities will complement the course material as much as possible.

There are 30 places, which will be open to both current and former students of the course. The OU will be sending out a mailing to next year's S381 students towards the end of 2002, and we will then begin registering students for the event. We have contacted all the members we know have been doing the course in 2002, and about a third of the places have been reserved. If we missed you, you can put your name down now. (Non-S381 students can also apply, but priority will be given to those who have studied the course if demand is high).

The cost of the event will be about £110 per head. This does not include overnight accommodation, but there is a Youth Hostel in nearby Hampstead Heath where you can get bed and breakfast for about £20, or there are local hotels for those who prefer greater "creature comforts".

On the Sunday, for those who stay overnight and those who live locally, we are hoping to visit one of London's tourist attractions (e.g. the Planetarium or the London Eye) and have lunch together.

The Radcliffe 24" twin refractor at ULO.

A Week at the British Antarctic Survey

by Iain J Coleman

MONDAY

It's a chilly winter's morning in Cambridge, and summer at the British Antarctic Survey (BAS). The canteen is only half-full, because a lot of staff are South for the summer season. Over the next couple of months they'll trickle back to HQ, until only a few dozen wintering staff are left on the Antarctic research stations. In my four years at BAS I've got used to this inversion of the seasons, though my wife still finds it odd that we go to midwinter parties in June.

After coffee, there's a meeting to discuss our plans to deploy optical instruments in Antarctica. We'll use these to make routine observations of the aurora, as part of our science mission to understand how magnetic interactions between the Sun and the Earth affect the upper atmosphere. Today the scientists and engineers are getting together to thrash out the various options, as a first step towards designing and building the kit. It eventually comes down to a choice between a simple, low-power system that we could make cheaply and deploy anywhere in Antarctica, or a more elaborate and powerful system that would have to be installed in one of the research stations. The first option gives us a lot of flexibility, the second gives us much better quality data. I argue that there's no point in building an instrument and sending it out to the most remote region of Earth if the data we get back isn't good enough to do useful science with. The rest of the scientists agree, and we go for the more powerful instrument. It's now up to the engineers to come up with a detailed design.

In the afternoon I get back to work on my own research project. The Sun's magnetic field is distorted when it comes into contact with the Earth's field: we call this "draping". There are various theoretical models of how this draping happens, and we think it's pretty important in determining where and when the magnetic fields interact. The funny thing is, there's been hardly any experimental work done to see if these models are any good or not. I'm trying to set that straight, by comparing magnetic field measurements from spacecraft upstream in the solar wind with the same measurements made by spacecraft nearer to Earth. This involves trawling through years of data from the NASA Goddard Space Flight Center, writing software to analyse it and produce pretty colour pictures, and then trying to figure out what it all means. That last bit has been eluding me for quite a while now.

TUESDAY

First thing today is a meeting with the other scientists in my team: Mervyn, the project leader, Mike, a senior radar scientist, and Gareth, my counterpart on the experimental side of things. I do the theoretical modelling work, and together we're trying to understand how the magnetic fields of the Earth and the Sun interact in a process called reconnection. We've all been working on the data from a particular day when the tell-tale signs of reconnection were seen at the same time by spacecraft at the edge of Earth's magnetic field, low-altitude meteorological satellites, and the ground-based radar network that includes the BAS radar at Halley Research Station. We're trying to tie all these observations together into one big picture, using my theoretical work to trace the magnetic connections between deep space and the upper atmosphere.

We've been battering away at this for a few months, and it's been getting pretty frustrating. We just can't get all the bits of the jigsaw to fit together well enough to convince ourselves, let alone the rest of the scientific community. Mervyn suggests a whole new approach, in which I fit my model to the spacecraft observations first, and then we take it from there. It's worth a go.

WEDNESDAY

I'm interrupted in my modelling work by an email from my brother, Jamie. He's in Glasgow, working in medical research, and he wants some advice. There's a mathematical model that he's trying to apply to blood vessels, but his forte is dissections rather than equations. Can I help?

It turns out that he's looking at exactly the same mathematics that some of the guys in my group have been using to explain the explosive release of energy in the Earth's magnetic field. I give Jamie some pointers, and I'm sure we'll talk more about it over some beers next time I'm north of the border.

Back to the space physics, and I've figured out that I can make sense of the spacecraft data if I assume that the Earth's magnetic field has been eroded away on the day in question, by a prolonged interval of reconnection. I meet up once more with Mervyn, Mike and Gareth, and show them my calculations. We decide that I should redo my modelling of the whole system on this basis, then if the results seem to make sense Gareth can use them to grind through the data once more.

THURSDAY

The big news today is about our main supply ship to Antarctica, the Ernest Shackleton. Since December, she's been unable to make it to Halley Research Station due to heavy ice. They've now decided to transfer the people and equipment from the ship to the base by aircraft, involving about 40 round trips of 600km. Sitting here in a warm Cambridge office, it's easy to forget the heroic efforts that the people down South have to make to keep our science projects running.

Today I'm organising my own travel for this year, which will be a bit less demanding. I'm off to the sunny south of France, for the European Geophysical Society's annual conference in Nice. It's a tough job, but somebody's got to do it.

Every Thursday afternoon, we have a seminar for the group, given either by one of us or by a visiting scientist. This week's speaker is Mark, one of our atmospheric scientists, who talks about the various instrumental effects on measurements of magnetic activity over the past 150 years. It sounds deathly dull, but it turns out to be an intriguing detective story, and a cautionary tale to anyone who uses long-term data sets.

In the evening, I spend some time on my second job. I've just started work as a tutor with the Open University, on a brand-new advanced astrophysics course. I did some tutoring as a postgrad in Glasgow, but since I joined BAS I've done no teaching at all. It's good to be a full-time researcher, without the added burden of having to prepare lecture courses and so on, but I do miss teaching students. It keeps your own understanding of the subject fresh and broad: without continually revisiting the basics, you quickly forget a terrifying amount. So that's why I've starting work with the OU: well, that and the cost of living in Cambridge.

FRIDAY

Some bad news today. I've just learned that a friend of mine has decided to quit research. He's been bouncing around Europe on various short-term postdoctoral contracts, most recently in France. Now he wants to settle down in France with his wife, but can't get a long-term research job there. So he's leaving science, reluctantly, and going into the computing industry. This is a damn shame, not just for him, but for the community as a whole: we're losing a first rate scientist, who made a unique contribution. He shouldn't have been forced to choose between his scientific career and stable family life.

This brings it home to me how well-off I am here at BAS. These guys are stuck with me till at least 2005, possibly longer. It's pretty rare for a young scientist to have such a long-term position in an academic institution, and most end up rushing from one-year contract to one-year contract. This means personal insecurity, and also makes it hard to do really long-term work.

Speaking of work, I finish the revised modelling and show it to the team. It all seems to make sense: Mike is quite happy with the match-up between my calculations and his radar data, Mervyn agrees that it all seems to hang together pretty well, and Gareth takes away the results to do some serious number-crunching on the polar electric fields. Rather him than me.

Leaving all that in Gareth's capable hands, I get back to my draping work. I've come up with a new way to analyse the data, and the results are starting to make some kind of sense at last. Just in time for the weekend.

The BAS Web site is at www.bas.ac.uk. Iain's article was first published on Science's Next Wave website (http://nextwave.sciencemag.org).

QUANTUM COMPUTERS - Part 3

In the third part of his series Andy Greentree explains the quantum of logic and quantum logic. > Part 1 > Part 2

Before we get on to quantum logic, we need to quickly revise classical logic. The fundamental entity of classical logic is the bit. This is a two-state system, for example it could be a switch that is on or off, a digit that is either 0 or 1 or a proposition that is true or false. In electronic applications involving logic it is usually designated as a voltage level, either 0V (logical low) or 5V (logical high). We could also consider a more abstract example where the bit is encoded by marbles rolling through a 'maze'. If the marble goes one way, we'll call it a 0, the other way, a 1. This isn't too crazy an idea, if we replace the marbles by a stream of electrons, then it looks just like conventional current flow in a circuit. Of course to make a useful computer we need more than just bits. We also need logic gates. Logic gates are devices which manipulate the bits - giving rise to useful computations.

In quantum mechanics, interference arose because of particles taking two (or more) paths to reach a common end point. So if we build the same 'maze' as we had before, but send quantum particles instead of marbles, then the particles can follow both paths simultaneously. If the particle can follow two paths at the same time, then the bit that it represents can take two values at the same time (it is in a superposition state much like the electron in the double slit experiments above). We call the quantum version of bits qubits.

Bits can be encoded by particles travelling along one path or the other. Qubits can go both ways simultaneously.

Having qubits instead of bits allows a whole new kind of logic. Because we associate bits with true and false, we can associate qubits with being true and false at the same time. The logic is called quantum logic and can be extremely powerful indeed. The fact that qubits can be in superposition states gives rise to the concept of quantum parallelism.

Given that you can go to any computer store and buy a 1.6 GHz PC with 512 Mbytes of onboard RAM, what do you need in a quantum computer before it matches up? Well the important thing to remember here is that any classical computer can only work with one number at any one time. A qubit in a superposition state can (in a sense) be two numbers simultaneously. Two qubits, can be 2*2=4 numbers simultaneously (00, 01, 10 and 11), three qubits 2*2*2=8 (000, 001, 010, 011, 100 101, 110, 111), and so on. If we could build a 300 qubit computer, then we could think of it as working on 2³⁰⁰=10⁹⁰ numbers at the same time. That's more numbers than there are atoms in the visible universe, clearly a powerful machine indeed.

Quantum Algorithms

The above discussion leaves out a number of important points though, not least of which is the outright speed of a quantum computer. In the above example its fairly clear that a 300 qubit machine working at even 1Hz (one operation per second) is going to outperform a classical computer, but what if our quantum computer is more modest? When does a quantum approach win out over the classical one and how do we understand this boundary? Such questions lead us nicely into the questions of computability and algorithmic efficiency. Without being rigorous, we need a way of deciding in a hardware independent way on the best approach to use to solving a particular problem. The reason for being hardware independent is that we know that machines will keep on getting better, so comparing an inefficient algorithm on a superfast computer with an efficient algorithm on a slow machine may not be a fair comparison. For algorithmic complexity we consider how many operations are needed to solve a particular problem and how that number of steps grows as the size of the problem grows.

To make this concrete let's consider the problem of finding an entry in an unsorted database. You might, for example, be presented with someone's telephone number and you need to find their name. Let's assume that the database has N entries in it. Without too much difficulty you can convince yourself that the best classical approach can be no better than starting at the beginning and working down the list, or maybe randomly trying entries. In either case, it will take (on average) N/2 searches until you find the correct entry.

Quantum mechanics yields a completely different way to perform the search and the algorithm is known as Grover's Search Algorithm after its creator, Lov Grover of AT&T labs. Grover's approach scales as √N. To show the importance of this kind of scaling, consider a database with all of the people living in London in it. There are roughly 6 million people in London, so the classical task of finding a match will take about 3 million searches. Quantum mechanically it will only take about 2,500 searches. If the database had everyone in the world in it, 6 billion people, the classical approach would take 3 billion searches, but the quantum only 80 thousand.

Grover's Search Algorithm works this way. First we take an input state and put it into a superposition of every possible state in our database. We then search the entire database in one step. The database 'tags' the appropriate state by rotating its phase through 180º. Then we interfere all the states. This makes an interferometer, more complicated than the electron double slit experiment, or the quantum 'maze' we showed above, but conceptually no different. Signal that has been put on the tagged state is not measurable yet, and so the database has to be searched more times yet, with interference after each step. It turns out that the number of searches of the database only grows as √N, which gives rise to the speedup discussed above.

Grover's Search Algorithm is only one of the algorithms being discussed at present, and it is not the one giving rise to the most excitement, that would probably be Shor's Factorisation Algorithm. Almost all of today's encryption techniques are based around public key encryption and the RSA algorithm. Without going into detail about how they work, what you do need to know is that their security is based on the fact that it is very difficult to factorise large numbers. What do we mean by difficult in this context? Well it is a long standing belief in mathematics that if you have an N digit number, it will take of order e^N operations to factor it.⁹ So supposing I have a 1024 digit number (similar to the one used in RSA) and I want to factorise it on my PC which is running at 1 GHz, I can do (about) 10¹⁷ operations per year. The universe is about 10 billion (10⁹) years old so if I could've started my calculation at the big bang, I would have done 10²⁶ calculations by now. e¹⁰²⁴ is about 10⁴⁰⁰, so I would have solved a tiny fraction of the problem. Quantum computers use quantum algorithms, and there is a quantum algorithm for factorisation. Instead of growing at e^N this algorithm only grows at N ³. This means that a 1GHz quantum computer could factorise a 1024 digit number in about 1 second. So if its going to be as fast to crack data as it is to surf the web, there will be a big shake up of how data is handled.

Lastly (and this was the context in which Feynmann discussed quantum computers in 1985) modelling quantum systems can be very hard on classical machines. If you look at quantum parallelism, consider that you need to model a quantum system on a classical computer, and you can see that some problems get very hard very quickly. If you can build a quantum computer, then some of these problems will be much easier, so successfully building quantum computers should prove a great boost to our understanding of quantum physics. This is more than just a self-fulfilling loop. Quantum physics is (according to our best investigations) the best description of the natural world, and we have no new theories that really seem to challenge it. So understanding quantum physics better is pretty important.

You now know some of the 'whys' of quantum computers, some of the fundamentals of quantum physics and you may even believe that quantum computers are inevitable given Moore's Law. In my next article we will look at some of the technologies that are being investigated to make quantum computers a reality.

⁹Note that no proof exists for this, it is entirely possible that there exists some classical routine waiting to be discovered that will factorise faster. If you find that algorithm you will probably become (a) very famous and (b) quite wealthy.

Further Reading: Quantum computers perk up with a cup of tea, Howard Baker, New Scientist, page 16, Feb 1, 1997. The last computer, Marcus Chown, New Scientist, page 26, Sept 2, 2000. The perennial classic on understanding physics has got to be this book; Mr Thompkins in paperback, George Gamow and Russell Stannard.

Andy has now moved on from the OU and taken up a job as Senior Research Associate in the Centre for Quantum Computer Technology, The University of New South Wales, Australia, where he is looking at silicon based approaches to quantum computing.

Leading Lights of Contemporary Science

Douglas Hofstadter - Cognitive Scientist

by Max Little

Douglas Hofstadter is considered by many to be one of the most original researchers in artificial intelligence and cognitive science today. His background is rooted in theoretical physics and mathematics, and in 1979 he published his first book, titled "Gödel, Escher, Bach: An Eternal Golden Braid" a strikingly original work that won him the Pullitzer Prize for non-fiction.

This book itself is quite strange. It contains chapters which are dialogues between fictional characters and more conventional explanatory sections. It broaches topics from particle physics, genetics to linguistics and formal axiomatic reasoning, and it does so in a refreshingly different style. It would be unfair to call it a textbook, but it has been said that it leaves many lay readers wondering exactly what he was trying to prove!

In this book, Hofstadter recounts how, as a teenager he read a book entitled "Gödel's Proof" which led to an interest in symbolic logic. He studied maths at Stanford University and started a physics doctorate at the University of Oregon in 1967. However, disillusioned with the way his physics career was going, he left Oregon and ended up teaching elementary physics to nurses in New York in 1973. Disliking this he returned to teach a course on Gödel's theorem and undecideability, and realising that he had to finish his PhD he took on a problem in solid-state physics in 1974, finishing it at the end of 1975.

His PhD subject was a theoretical investigation into the allowed energies of electrons in a crystal subject to a magnetic field. The topic was to combine the results of treatments of electrons in a crystal and in a homogeneous magnetic field which up till then had been calculated separately, and the solution required a continued fraction - a recursive structure. When the ratio of electron time periods in the lattice and the field separately is rational, e.g. q/p, then there are exactly q energy bands, but when irrational the bands shrink to infinitely many points drawn from the Cantor set - another recursive, fractally nested structure.

He then goes on to liken the grammatical structures in ordinary spoken language to the "recursive grammar" of Feynmann diagrams showing renormalization in particle physics, whereby all interacting particles are defined in a nested way, with electrons creating virtual photons which in turn create virtual electrons and so on ad infinitum.

This analogy is followed by other analogies to exact mathematical recursion found in the formal computer sciences started by Gödel, Church and Turing and depicted in works of the famous 20th-century artist M.C. Escher. He also shows how such structures can be found in the melodies of the preludes and fugues of Bach. This leads to discussions about how other isomorphisms, or maps, to and from fractal structures can be found all over the natural world and in the artefacts of human cultures. The goal of this rambling tome was to show how fractals, or recursive structures which he calls "strange loops" are essential to the study of conscious mental behaviour, and at the same time define some of its more baffling consequences.

After producing this book, Hofstadter continued his research into models of mental cognition at Indiana University. Some of his latest research shows how it is possible for a computer program, given one stylised letter of the alphabet, to generate the rest of the alphabet in that same style - a typographical design trick usually only associated with considerable human creative skill and mature aesthetic appreciation. Other research includes the relationship between words and concepts, the mechanisms underlying human error-making, and the mechanisms underlying discovery and invention in geometric and symbolic mathematics.

He recently published a book outlining the results his latest research as head of the Fluid Analogies Group, covering topics such as the recall of old analogies and the formation of new ones in creative thinking, the discovery of rules behind mathematical sequences and the "unconscious juggling of mental objects".

Having contributed a column for the Scientific American called "Metamagical Themas", a twist on the column that preceded it called "Mathematical Games", written several articles about Chopin and penned translation in verse of Pushkin's Onegin, from this and other works it is easy to see that Hofstadter is a highly unconventional scientist! Despite this, his biggest contributions to the field appear to arise partly from the scope, originality and creativity of his scientific and mathematical knowledge. For more information, visit his website at www.psych.indiana.edu/people/homepages/hofstadter.html.

How To Get A PhD Studentship

by Paul Ruffle

For many of us, getting an OU degree in physics is a long drawn out process, especially if you stick to doing just one physics related course per year. However, there comes a time when you have only one or two courses to do, and it dawns on you that you can give some serious thought as to what to do after that very last exam. If, like me, you find that the kids have grown up and left home, you may want to consider seeking a PhD research studentship in some aspect of physics that particularly interests you.

Now this all might sound a bit pie in the sky. It did to me, but two years ago I attended YPC2000 in Chester (organised by Nexus) and found out that doing a PhD was not as out of my reach as I had thought. Firstly, the funding from the research councils (PPARC and EPSRC) has gone up considerably over the last few years and currently stands at £9,000 per annum (tax free), with further increases over the next few years agreed by the Government. Tuition fees are also paid on your behalf and expenses for attending conferences or travelling abroad to collect data are usually covered as well.

Secondly, the attitude of many Physics Departments in the UK is very positive to older, OU graduates. As a post doc in my group said to me the other day, "Give me a choice between a regular student and an OU graduate and I'll pick the OU graduate every time - they have got their degree while working or bringing up a family and they did it because they really wanted to!" Not that every University has such an enlightened attitude. One or two that I contacted were decidedly 'sniffy' when it came to taking my application seriously.

So, how do you go about it, what are the requirements and what is involved? Well, it means working for three years (full-time) researching a specialist area of physics within a University department and producing a unique piece of work (your thesis), that contributes to the total body of scientific knowledge. In practice you spend the first year learning your subject and the 'tools of your trade' (computer analysis tools, laboratory equipment, etc.), the second year doing the real work and the last year writing up your thesis.

To start with you need a first or upper second honours degree to get research council funding. That means averaging grade two passes on your best 120 points at second and third level. Obviously your third level grades count for more, but you can afford to have one duff grade. The OU publish a detailed matrix of how honours classifications are arrived at.

Given the somewhat arbitrary nature of the way post graduate admissions are administered - i.e. totally differently at each University - you need to apply early and to as many relevant departments that you can identify. For example there are around 40 physics departments in the UK, but in my case I only identified 27 groups doing astrophysics. By early, I mean start trawling each University's Web site in December and get your applications sent off in January. I did not start doing this until February and missed out on being considered by at least six Universities. Interestingly, my offer came from a University that I had inadvertently missed from my original list - so be thorough!

As part of your application you will need two academic referees. Ideally you want OU tutors with academic standing that actually know you and have a feel for your academic potential beyond the bare numbers on your OU record. Let them know what you are planning and ask nicely so that they can draft your reference on their PC and then print it out numerous times for all those different applications you are going to make.

Seek advice and guidance from as many people as you can. You will be surprised at how helpful and encouraging people can be, as a result of an initial email enquiry. Write yourself a two page PhD application resume to include with your application. Give a little background about yourself and how you have got to this point in your life of wanting to do a PhD. Remember that younger applicants have not got your wealth of experience, so use that to sell yourself. In fact, in the first instance I emailed such a resume to everyone on my list.

If you cast your net wide, you should get several interviews. They will all be quite different, some very informal, others more formal (I had one that was more like a court martial). Beforehand read up on what the group is researching (from their Web site) and show interest and enthusiasm in what they are doing. Usually you will be passed round the members of the group and each will tell you about their research interests. This way you can pick up a lot of useful background knowledge for your next interview!

If you can't move to a new location, you need to get to know your local University physics department so that you can get a 'foot in the door'. Attend any lectures or colloquia that you can and make a point of introducing yourself to the staff. Tell them of your plans and seek their advice. To show willing, volunteer for any unpaid lab work that might be on offer. You could also consider doing an MSc locally (part or full-time) and improve your chances of getting a subsequent PhD studentship.

Winning a funded studentship is a privilege (it will cost the taxpayer over £40,000) and a wonderful opportunity to be part of cutting edge physics research. If you really want to do it and are determined - you will!

Useful Web sites: www.pparc.ac.uk, www.epsrc.ac.uk, www.findaphd.com.

Paul has just completed his OU degree with S381 - The Energetic Universe and has now started a PPARC funded three year post graduate PhD studentship in astrophysics at UMIST in Manchester.