The most critical dimensions of an ICP torch are the distance between the outer gas inlet and the top of the intermediate tube, the magnitude of the space left between the two outermost tubes and the overall torch diameter. The reduction in the torch length while keeping constant the critical dimensions allows injecting directly the aerosol generated by a conventional pneumatic concentric micronebulizer into the plasma by simply fitting it at the torch base. This approach is presented for the first time in this paper. A MicroMist has been used in conjunction with the reduced length torch. The new system has been evaluated in terms of ICP-AES sensitivity, signal stability and matrix effects caused by nitric acid. The results have been compared against those provided by a double pass, a low inner volume cyclonic spray chamber (Cinnabar), a torch integrated sample introduction system (TISIS) and a direct injection high efficiency nebulizer (DIHEN). When using the reduced length torch, at the optimum nebulizer liquid and gas flow rates in terms of sensitivity (i.e., 60 μl min−1 and 0.15 l min−1, respectively) the MicroMist generated aerosols with droplets whose diameters were even higher than 100 μm. Therefore, the selection of this nebulizer was based on criteria other than signal stability such as robustness and lack of tip blocking problems. Despite of this, the results were not as different from those afforded by a DIHEN as expected. For the reduced length torch, the maximum signal improvement factor with respect to a double pass spray chamber was from 1.5 to about 3 depending on the liquid flow rate and the emission line. Similar limits of detection were found for both systems. Meanwhile, these ranges were from about 1.5 to 4 and from 1.3 to 7 for the TISIS and DIHEN, respectively. Note that under its optimum operating conditions, the DIHEN generated droplets with diameters above 30 μm. As expected, at low liquid flow rates, wash out times were more than 20 times shorter for the reduced length torch than for a double pass one. As regards the matrix effects, it was found that the two direct sample injection systems tested behaved alike, the interferences becoming more severe as the delivery flow rate went up. A problem found with the MicroMist associated to the reduced length torch was that the signal stability was poorer than for the remaining systems.