This work presents a simple method for obtaining homogeneous sample surfaces in matrix-assisted laser desorption/ionisation time-of-flight mass spectrometry (MALDI-TOFMS) for the automated analysis of peptides and proteins. The sample preparation method is based on applying the sample/matrix mixture onto a pre-deposited highly diluted matrix spot. The pre-deposited crystals act as seeds for the new sample containing crystals which become much smaller in size and more evenly distributed than with conventional methods. This 'seed-layer' method was developed, optimised and compared with the dried-droplet method using peptides and proteins in the 1000-20,000 Da range. The seed-layer method increases the surface homogeneity, spot to spot reproducibility and sample washability as compared with the commonly used dried-droplet method. This methodology is applicable to alpha-cyanohydroxycinnamic acid, sinapinic acid and ferulic acid, which all form homogeneous crystal surfaces. Within-spot variation and between-spot variation was investigated using statistics at a 95% confidence level (n = 36). The statistical values were generated from more than 5000 data points collected from 500 spectra. More than 90% of the sample locations results in high intensity spectra with relatively low standard deviations (RSDs). Typically obtained data showed an RSD of 19-35% within a sample spot as well as in-between spots for proteins, and an RSD of < or = 50% for peptides. Linear calibration curves were obtained within one order of magnitude using internal calibration with a point-RSD of 3% (n = 10). The sample homogeneity allows mass spectra (average of 16 laser shots) to be obtained on each individual sample within 15 sec, whereby a 100 spot target plate can be run in 25 min. High density target plates using the seed-layer method were prepared by spotting approximately 100 picoliter droplets onto the target, resulting in sample spots < or = 500 microns in diameter using a flow-through piezo-electric micro-dispenser. By using this automated sample preparation step lower standard deviations are obtained in comparison to manually prepared samples.