Vertical loops or modified vertical loops are basically frictionless springs which are used for canine and anterior tooth retraction. The design and selection of a proper loop or retraction spring should be based on a number of scientific criteria. Foremost among these would be a sufficiently high moment-to-force ratio so that root apices are not displaced mesially or anteriorly. A retraction spring with zero angulation of its horizontal-occlusal arms delivers a moment when activated to produce a force. The ratio of this moment and force is constant throughout the elastic range of activation of the spring. The higher the moment-to-force ratio, the greater is the clinician's control over the apices of the anterior teeth. An analysis of design factors demonstrates that the higher the loop occluso-gingivally, the shorter its horizontal length occlusally, and the greater the gingival horizontal length as in a T loop; these are significant factors in increasing the moment-to-force ratio. The placement of helices is a useful design consideration but the main effect is in reducing the load-deflection rate. By keeping these design factors in mind, the clinician can build into his retraction springs, without the placement of any gable bend, the largest possible moment-to-force ratio so as to optimize his tooth movement. Although it may be possible to design retraction springs to deliver an adequate moment-to-force ratio for controlled tipping around the apex of an incisor or a canine, translatory movements are not possible, considering the intraoral limitations on spring height. This can be overcome by the placement of gable bends or angulation in a vertical loop or retraction spring. Unfortunately, with the typically used high-load-deflection-rate vertical loops, activation to achieve the desired moment-to-force ratio is too critical, exacting, and changeable with small displaced movements of the tooth. This can be partly overcome by utilizing designs that have not only the highest possible moment-to-force ratio during pure horizontal activation of their arms but low-load deflection rates as well. Because of the low load-deflection rate, moment-to-force ratios are relatively more constant if a gable bend (angulation) is placed. The science of spring design as applied to the problems of canine and anterior tooth retraction in this article allows the clinician to optimize the design of his retraction springs. More important, with properly designed springs, it allows him to estimate with relative accuracy the force systems produced and to avoid undesirable side effects which might not have been apparent from superficial observation.