It's that time of year, the undergraduates are finished and wondering what to do WITH THE REST OF THEIR LIVES OH NOES! So a number of them have been directed to me, and told "ask her what it's like, see if she'd recommend it". I shall restate most of my advice below, for those of you who haven't got to hear it from my face and because, apparently, it's not bad advice.

The first and biggest question: Why do you want to do a PhD?
It's the first thing I ask them, often it's one of the key questions in a PhD interview and, well, it's good to see if they have a clue what they'd be letting themselves in for (undergrad final year projects are generally carefully planned to be manageable, PhD projects, not so much).  I've had a number of responses, the most common being a shrug of the shoulders and an "I'm not sure, I just thought I'd like to do one". To a lesser extent I've heard "I don't want to get a boring job" and "I'd like to be a lecturer".
The last two are easiest to address, then we'll deal with the "don't knows". Lecturing postitions aren't the easiest to come by these days. In biotech only 5% of PhD graduates have a chance of getting to be lecturers while in engineering it's a bit closer to 20%, but still, not a guaranteed endpoint (stats off the top of my head and an engineering colleague's, handy confirmatory link found by @alanmrice). As for I don't want a boring job, why don't you want a boring job? First off, real world jobs are more likely to come with lovely things like stability and a pension.  Secondly, haven't you seen me repeat the same experiments for weeks/months? That's what I'll be doing for years, there's some scope for thinking outside the box and planning my own work, but most of it is a slog through identifying problems and possible solutions, over and over.
If you don't know why you want to do a PhD, then really and truly think hard and find a reason.  If you can't then don't do one, get a job, do a masters, travel for a year, volunteer for some charity you've always admired, do anything more worthwhile than starting a PhD that you don't particularly want. At the very least talk to a lot of PhD students, and then talk to a lot of people who supervise groups you'd be very interested in working with (note the word very, a vague interest in a broad field might not be enough to drag you through 5 years of drudgery).

Next big question I ask them: Do you actually need a PhD?
Well? Do you? My first piece of homework for my young advice takers is "go look up the jobs websites, see if there's any jobs you'd like, and see what qualifications you need for them".  If the job you want most in this world doesn't need a PhD, then why waste years getting a qualification that you don't need to have when you could be earning far more money and doing the job you always wanted.  There are jobs that having a PhD could overqualify you for. If you think like a hiring manager then you'd think "this person has a great degree and is probably only applying for this job until they can find a better one, I won't be able to keep them".  Sometimes a relevant masters is needed for the job, a masters will only take a year or two, and will let you enter the jobs market sooner.  The final salary difference between a masters graduate and a PhD graduate makes you wonder if the PhD is even worth the extra effort.
There are many jobs that require a PhD, but the eagle eyed among you will have noticed that many also accept a larger number of years of experience outside a PhD in its place.  Prior to starting this PhD, I figured I'd like an industrial research gig, but the job ads at the time all wanted a PhD or ten years experience, so I decided a PhD would be quicker to get than ten years in industry.
A PhD is basically shorthand for "I have 3/4/5 years experience, learning new skills and (mostly) managing my own work".

The big question for me (posed by either students or the people who sent them) is Would you recommend doing a PhD?
Ehhhh, thanks for the loaded question there Mr. Supervisor Sir.
Yes and no, I'd recommend a PhD, but they're not for everyone, and they're certainly more challenging than you expected.  The plus sides, I get to work on a project that I play a large part in directing and planning, I can basically choose my own working hours, and have a lot of control over the budget for my work, I work with some very smart people on science that's fascinating and in a growing field, I get to work with people in different fields and bring physics and biology a little closer. On the downsides, not everything works out and I spend a lot of my time troubleshooting and optimising experiments, not all the expertise I need is available for questioning so I have to find a lot of my own answers, the wages aren't great and the hours can be long, I don't get to pay income tax or PRSI (which sounds good, but I lose out on state benefits). Of course, these are just my experience, some people work in older more defined fields, some people play a minor role in planning (and certainly don't hold the budget codes), some people's experiments work out perfectly with appropriate controls first time (almost) everytime. The best thing to do is to ask a lot of PhD students questions and average the answers (or try to control for supervisors/institutes/students themselves).
I like learning new skills, but it takes time from the project which can cause different problems (also, you need to find a person to learn the skills from). I enjoy teaching, demonstrating to undergrads is a compulsory part of the job and it can be a mix of incredibly frustrating and very rewarding, even within the same classgroup (or ten minute length of time).  I don't like the regular report writing, six monthly progress forms just depress me. I don't like the coffee on campus, but that probably shouldn't be the main thing that swings your decision...

Well, it seems interesting, where do I start looking?
The standard answer is to look up papers in your field of interest and start hassling lead investigators (usually the last name listed on the paper, in Biology anyway). I'd highly recommend talking to your final year project supervisor, or another member of the department who's familiar with the field you want to get in to, quite often people have contacts they can share with you, or can advise you to maybe consider a different potential project supervisor.
In addition, this person can often advise you on where to look for funding. Some investigators will have their own funding to take on a PhD student but some will not.  Departmental newsletters often contain lists of funding opportunities and deadlines, so ask about them. There's always the possibilty to apply for PhD specific grants with your potential supervisor from various sources such as state agencies (IRC would be the biggest such funder in Ireland), charities and private industry.  In fact, if you end up getting one of those jobs I recommended looking for earlier, you might find that your employer will help to fund you to do a PhD that would benefit both you and them.

The supervisor I really really really want to work with has no money and I didn't get a scholarship, so I'll just do it for free and see how that goes.
NO NO NO NO NO NO. A fulltime PhD is a fulltime job with guaranteed overtime and you'll end up spending most of your waking hours thinkng about it.  If you came across someone working a fulltime job for no/low wages, you'd report their boss for exploitation, why should it be any different for a PhD researcher? Why would you want to work for someone who would let you work without wages for so long? I have spoken to people who would refuse to take on an unpaid PhD student, no matter how much the student wanted the position, because students don't deserve to be treated like that.
Sure, you did your undergraduate degree unpaid, what's three more years? Well, for a start, it may not be three years, or four... The idea that you could work parttime isn't going to cut it, maybe if you were doing a parttime PhD, working would be feasible, but not full time.  It takes too much of your life, sleep and energy.  If you can't find funding with the person you want to work with most in the world, find another person, or see if there's any forthcoming grants that you could be squeezed in on, but absolutely do not do a full time PhD unpaid.

So, in summary:

  1. Figure out why you want to do a PhD
  2. Figure out if you really need it, and would you be better off getting a masters or applying directly
  3. Ask your final year supervisor and other deparmental folks about where to start looking and where the money can be found
  4. DO NOT DO AN UNPAID PHD, you're worth more than that

And feel free to ask me any more questions, I've bucketloads more of this sort of thing in my head!