Observations of the escape responses of Euphausia superba and Euphausia crystallorophias under pack ice (O'Brien, 1987) and of Nyctiphanes australis in southeast Tasmania (unpublished observations) are discussed together with previously published data on the escape reactions of these and other euphausiids. Similarities in the form of the escape response for surface-swarming species of euphausiids are apparent. It is proposed that there are three levels of escape response whose manifestation is primarily dependant upon the aggregative state of the prey and the size and proximity of the predator. The adaptive significance of the three levels of response is discussed in relation to predator evasion. Central to the study of the distribution and regional variations in the density of euphausiid species and to the success of commercial trawling operations is the efficiency of the locating and sampling equipment, especially since krill may actively avoid nets (Marr, 1962; Pavlov, 1969; Sameoto, 1980; Everson, 1983) and remote camera equipment (Kils, 1981). As stated by Hamner (1984) in reference to E. superba, a knowledge of the avoidance or escape responses and the general behavioral biology of krill is of fundamental importance to the design of sampling gear, harvesting strategies, and also to the evolution of its predators, e.g., whales. Published accounts of the escape responses of euphausiids are scarce and are restricted to direct observations of surface-swarming species such as Euphausia superba (Pavlov, 1969; Ragulin, 1969; Hamner, 1984), Euphausia vallentini (Ragulin, 1969), Euphausia pacifica (Hanamura et al., 1984), and Meganyctiphanes norvegica (Nicol and O'Dor, 1985) or indirect evidence from the use of remote cameras and echo sounders (Sameoto, 1980; Kils, 1981). In the present paper further direct observations of three species of surface-swarming euphausiids, E. superba, Euphausia crystallorophias, and Nyctiphanes australis are described in relation to the previously published data, and the general features of the escape response are discussed for the group as a whole. In the present context the term swarm describes any cohesive aggregation of krill, even schools when there is no purpose in distinguishing between the two terms. The term school refers to a cohesive group of krill in which all the individuals are polarized, i.e., show parallel orientation. With both terms there are no restraints on density or dimensions. All unacknowledged references to E. superba, E. crystallorophias, and N. australis are taken from O'Brien (1987), or are unpublished observations made by the author.