In addition to the points already stated, it is important to consider whether you are doing amperometric or potentiometric measurements.  In potentiometric studies, I would always use a reference electrode with a salt-bridge junction to avoid interactions with the calomel or silver chloride reference.  If using an electrode without a liquid junction, such as a pH electrode or ISE, then the sample solution must be well poised with the reference ion, i.e., a pH buffer or high concentration of the selected ion.  In amperometry, always use a three electrode system to avoid concentration polarization effects that would change the potential of the reference electrode.

I think you should really read A. Bards first chapter:

http://www.amazon.com/Electrochemical-Methods-Fundamentals-Allen-Bard/dp/0471043729

I will just want to add an important point. The nature of your system plays a vital role. For example Ag/AgCl or Calomel (SCE) are a great choice for RE when you are doing experiments in aqueous environments/solutions. BUT if you are using a mixture lets say ACN:Water or a non-aqueous  solution the fact that the RE electrode contains an inner aqueous solution will cause a big resistant in the RE-Solution junction. Some Commercially available electrodes have a membrane that is sensitive to organic solutions. Then your choices can be Ag/AgCl wire (quasi RE) or Ag/Ag+ or Ag/AgBr       

 Ag/ Ag+ can be used. 0.01 M AgNO3,AgClO4 or other silver salt can be  chosen

electrolyte 0.1 M TBAP/CH3CN)  or DMF or DMSO or PC can be used.

I really like leakless RE, good for HF and most solutions.

http://2in.com/products/reference/reference.html

The link that Julia provided is very interesting, although it lacks real scientific information about the nature of the junction. Indeed, it goes against everything I thought I knew about electrode junctions and how they function. If true, it is a real scientific breakthrough. I will reserve judgement on this until I am able to get more details about the composition of the junction and its characteristics. If anyone can provide more information on this device, please let me know.

"Leak-free" may be a commercial brand for a "diffusion" type juction:

http://www.vl-pc.com/default/index.cfm/technical-info/reference-electrodes

Thank you, Cornel, for your suggestion that this is actually a diffusion-type junction. However, their description of the "leak-free" junction (copied below from their website) does not at all sound like a diffusion junction. A diffusion junction cannot be described as not leaking at all since the junction electrolyte is able to diffuse between the sample solution and inner reference electrode solution. Also, high conductivity is not a usual characteristic of a diffusion junction. In fact, if you read through the description, you will see several other characteristics that distinguish this junction from a diffusion type.

No, your explanation cannot account for the leak-free junction described in their website. I still hope that a more satisfying explanation will be presented by someone having detailed technical information. But thanks again for your suggestion.

"Our new Leak-Free junction series, LF, enables you to perform your experiments in organic solvents, perchlorate and silver salts solutions, or Hydrofluoric acids without being worried about clogging or degradation! It can also be used for long term experiments without the worry that the filling electrolyte be diluted or run out. This Leaking-Free reference electrode uses our newly developed conductive junction (patent pending). The filling electrolyte is confined to the barrel and will not leak at all (zero leakage). The Junction has very high conductivity with resistance under 10 K-Ohm. It has exceptional mechanical stability, zero swelling, resistance to organic solvents, and robust. The junction potential is independent of the sample nature or ionic strength. The electrode body is constructed from PEEK for superior chemical resistance. The filling electrolyte (3.4 M KCl) does not leak through the junction which prevents sample contamination with chloride and potassium ions. This means no clogging and no need for double junction. Since the electrode construction does not involve any glass, it can be used in Hydrofluoric Acid solutions."

Thank you, Richard, for the comment.

A 10 K resistance is at the high limit accepted for a reference electrode. Thus the junction doesn't has " very high conductivity ". I have seen the picture of the reported LF electhode (small electrode diameter, without a filling hole, possible a gelled electrolyte inside, relatively high area for the junction).

" Leak Free reference electrodes utilize a unique junction that is highly conductive (< 10 kΩ) but not porous. There is no solution migrating through the junction in either direction making the reference totally leak free! " from

https://www.warneronline.com/product_info.cfm?ID=367

Cornel, I agree that even though they claim that the junction has "very high conductivity with resistance under 10 K-Ohm", this is not high conductivity. The fact that they also claim that, "The filling electrolyte (3.4 M KCl) does not leak through the junction which prevents sample contamination with chloride and potassium ions.", indicates to me that diffusion cannot be taking place. There has to be another explanation. Any more thoughts? Thank you.

Osmosis? Semipermeable membrane for the junction? The osmotic pressure drives the solvent from the sample through the junction (sample is not contaminated, but the electrode is): this may explain why they recommend to keep the electrode in distilled water (not a saline solution as for a common reference electrode).

That's another good idea, but it suffers from at least two problems. First, ions have to be able to transfer across the junction in order to maintain charge neutrality and, if the solvent is the only substance that can transfer across the semipermeable membrane, electroneutrality cannot be maintained. Second, as you pointed out, the body of the reference electrode is sealed so that, if solvent transfers into the reference electrode compartment, eventually the back pressure will prevent further osmosis.