It might be that you don't have protein anymore. What happens if you change your crystals from the 'old' drop to a fresh one? Another possibility could be that the protein becomes oxidised. Finally, among other possibilities, I would suggest to check if the protein remaining (if any) in your drop is functional / folded properly / not degraded.

Crystals stop growing because the conditions reach an equilibrium which may be due to various things.  I agree with Leonard's answer above.  Also check to see if a film (of cross-linked or denatured protein or whatever) is formed on the drop surface preventing water evaporation.  Even if your protein in the bulk solution is still plenty and in its crystallizable form, an equilibrium is established if the drop cannot get concentrated further (due to the presence of the film).  In that case, just cutting open or removing the film with a tool may be enough to get them grow further for a day or two.  If your protein contains exposed Cys, it may cross-linked.  Take the solution for native-PAGE or even SDS-PAGE with and without thiol-reducing agent in the loading dye to see if you may spot any sign of cross-linking.  Sometimes, the protein in the crystals may be cross-linked, as seen by their rubber-like behavior - they do not break but would bend upon getting poked by a tool.  If this is the case, consider addition of a thiol-reducing agent in the experiment (and using the vitamin E-containing oil or replenishing the agent if needed).  Another possibility is that irregularity/imperfect molecular arrangement is formed such that deposition of more molecules are prevented.  This may be the case if different crystals stop growing at different time (ie. smaller ones can still grow further while bigger ones stop growing when they reach a certain size) and usually when they reach the same sizes.  Could result from various factors.  Check for vibration - add extra things (sponge, cotton pad, other mechanical gadget, etc) to damp it.  Check for oligomer formation , dusts, ... Was the protein purified through gel filtration column?  Did you filter the protein sample before setting up crystallization drops?  Try setting bigger drop size with the same and different protein:precipitant volume ratios.

Mrs/Miss Chitnumsub,

Nevertheless are your process is exo- or endothermic one, the partial pressure of the product (PPP) at so-called "source" meaning matter solution (or gas phase employing the vapour growth techniques) should be higher that this one of the growing crystal (PPGC). The perturbation of flow rate yielding the source composition has led to decreasing in mentioned above PPP and thus to decrease in the value below the growing crystal one. In such as cases the total chemical potential, defining by-RTln(PPP/PPGC) has become positive and thus a crystal growing process - unfavourable one. In this respect you could apply P in a gradient scheme, keeping the system isothermic as it is according your explanation. Have you any information about the thermodynamics of the transport reaction/s or about the growth reaction/s? Because of using gradient P-approach, there has possibility to obtain a fine polycrystalline sample object, than a good quality single crystal. Another posibility is to use gradien T-aproach towards, however, one of the components of the "source" composition.

Did you check for bacterial contamination?  This would produce the "precipitate" and destroy the dissolved protein.  For reasons explained by the above responses the dissolved protein concentration may not have remained adequate for further crystal growth.  There are other methods that can counter this issue:

P. Todd, S. K. Sikdar, C. Walker and Z. R. Korszun.  Application of osmotic dewatering to the controlled crystallization of biological macromolecules and organic compounds. J. Crystal Growth 110, 283-292 (1990).

C.-Y. Lee, M. G. Sportiello, S. P. Cape, S. Ferree,  P. Todd, C. E. Kundrot, S. Lietzke and C. Barnes.  Characterization and application of osmotic dewatering to the crystallization of oligonucleotides.  Biotech. Progress 13, 77-81 (1997).

M. Sugiyama, S. Sengupta, P. Todd and V. H. Barocas.  Concentration control for protein crystallization via a continuously-fed crystallization chamber. Lab on a Chip 8, 1398-1401 (2008).