Can we fire a single femtosecond pulse of light without an associated train of additional pulses? What sort of equipment allows for this? What sort of mode-locked laser systems allow for the production of single isolated pulse of light (as opposed to a train of more than one pulse), where the individual pulse duration is on order picoseconds ($\approx 10^{-12}$ seconds) to femtoseconds ($\approx 10^{-15}$ seconds)?  Is there any reason I couldn't use a mode-locked Ti-sapphire laser to produce a single pulse of some desired / achievable duration in the femtosecond regime?  Or is it generally challenging to drop the pulse repetition rate to a sufficiently small frequency to allow for shutters / etc. to filter out the rest of a pulse train?
My guess is that if we can get the pulse repetition rate down to the MHz regime, we can start to think about diverting or blocking additional pulses using voltage or mechanically switchable filters?  What do people actually use?
I should also mention that I'm particularly interested in emission wavelengths in the range of something like $100 \space nm$ to $1.5 \space \mu m$ or so.
...
In terms of the different types of femtosecond lasers, there are fiber lasers, bulk lasers, dye lasers, semiconductor lasers, even free-electron lasers that one can use.  Supposing I need something like a few hundred milliwatts to a watt of output power, and I'm interested in the visible to near-visible wavelength range, are there any families of femtosecond lasers particularly suited to the above task?
 A: Typically for the wavelength band, power requirements and pulse duration you mention, Ti:Sapphire lasers are employed. AFAIK it's not uncommon to use the electro-optically shuttered output of a regeneratively mode-locked Ti:Sapphire as a seed for an ultrafast regenerative amplifier.
For example, a Spectra-Physics Spitfire Ace regenerative amplifier can be seeded using a Mai Tai SP oscillator. The oscillator has a mode-locking rep rate of 84 MHz, but the regenerative amplifier has a maximum repetition rate on the order of several KHz, so presumably some shuttering is done to pick out single pulses. In this manner, pulses in the 100 femtosecond regime can be created, with reasonably high power output.
For shorter pulse generation ($<20$ fs), these pulses mentioned above can be used in a non-collinear phase-matched optical parametric amplifier with subsequent pulse-compression using fused silica prisms.
A: In order to isolate one pulse out of a train of pulses in the MHz regime a Pockels' cell is generally used in commercial Ti:sapphire laser technology. This allows for single-pulse picking and amplification and can be used with most femtosecond laser oscillators. It is possible to cavity-dump a femtosecond oscillator and lower the repetition rate that way, but for a single pulse using an external Pockels' cell is the most straightforward way. Wavelengths in the range of ~500 nm up to 1600 nm can be achieved relatively simply through optical parametric amplification, and multi-stage nonlinear conversion can extend that range into the UV and up to around 20 microns at the cost of conversion efficiency. Depending on your need (wavelength, pulse energy, and/or average power) there are many different solutions, ranging from a simple Kerr-modelocked Ti:sapphire oscillator that can produce a few nJ per pulse at 100 MHz (200-500 mW average power) to multistage amplifiers that produce mJ pulses in the 1-10 kHz repetition rate range (~1-15 W). Amplified pulses can in turn be used to drive optical parametric amplifiers (OPAs) that generate tunable radiation (typically in the microjoule pulse energy range). 
Hope this helps!
