What is the max. permissible tubing(0.3mm)length, HPLC system:shimadzu model, that could be used to make connections [ injector-column-detector] so that extracolumn effects can be kept to a minimum?
Thank you for the assistance.

The distance from injector to column is not as critical as the distance from column to detector. From column to detector should always be 'as short as possible'. If you have use a longer piece to reach your detector, put it between the injector and column. Also I would suggest going to 0.005" id tubing to minimize the ECE even more.

The statement on the tubing by Merlin applies only to gradients. For isocratic runs, if makes no difference where the tubing is. How small it needs to be depends on the column that you want to use.

The previous statment is not quite correct. Too much volume between the column and detector leads to increased band spreading, even during an isocratic run.

If you are using gradient, you will definitely want to minimize the TOTAL system volume from the pump's gradient valves to the column, and that volume usually comes mostly from the solvent mixing system rather than the tubing. Also, I believe the autoinjector can make a significant contribution that would apply to both isocratic and gradient. I doubt that your connections from one component to another would have much of an impact relative to the impact within a single component.

We need to clarify, since everybody talks about different issues.
There are two effects that are important: extra-column bandspreading, and gradient delay volume. Of course, if one does not run gradients, the gradient delay volume is not relevant.
Extra-column band spreading are all the effects that make the peak wider than it were, if it were exposed to the column alone. As components behind the column, this includes the detector cell, and the tubing to the detector cell. These components increase the peak width from the column no matter if you are running a gradient or isocratically.
Then there is pre-column bandspreading. If you are running a gradient, your precolumn bandspreading can be nearly as large as the United States without affecting your chromatogram. This means large injection volumes are possible due to the fact that the sample enriches at the column top. If you are working isocratically, all components from injector to the column widen the peak. You can either make the connection shorter and the volume smaller, or you can play tricks such as dissolving the sample in a weaker solvent than the mobile phase. All of these things reduce the bandspreading.
The other factor is the gradient delay volume. This is due to the fact that low-pressure gradients are mixed signifcantly upstream of the column. Therefore it takes time until the gradient hits the column, and the separation runs isocratically for a while. On one hand, this is a waste of time. On the other hand, this causes trouble if you are specifying retention times and transfer the method from one system to another. There is a fix: the most modern low-pressure gradient systems have the ability to make a delayed injection, just at the moment when the gradient hits the column. This allows an overall reduction in run time and makes the gradien reproducible from system to system. Of course, one can also use high-pressure mixing with two pumps to get around this problem.
Hope this clarified things.