Background: Clinicians frequently experience difficulty in eliciting the reflexes of elderly patients using standard methods due to paratonia/frontal rigidity. If reflexes are incorrectly thought to be absent, important diagnostic errors may be made. Neurologists use alternative methods when technical difficulties require them, but these are not widely used by non-neurologists. Methods: A neurologist and a medical student both used standard and non-standard techniques to assess reflexes of the lower limb in geriatric inpatients, aged over 65, to determine which method permitted the most confident assessment of the presence of knee and ankle reflexes. Results: 45 patients were assessed. The consultant found that in 20 patients (44%) all three knee reflex methods examined produced similar results. When the methods produced different results, the “superior patellar supine” method was the best single method overall (best or equal best in 19 patients (42%)). For the ankle reflex all four reflex methods examined produced similar results in only 7 patients (16%). When the methods produced different results the “Achilles strike elevated” method was best or equal best in 32 patients (71%) and the “plantar strike” method in 29 patients (64%). If the student had relied on standard methods alone, reflexes would have been incorrectly called absent in 28 limbs (37%) for knee jerks and 52 limbs (84%) for ankle jerks. Supplementing standard methods with alternative methods reduced these error rates to 19% and 21% respectively. Conclusions: Our findings indicate that a reasonable practical approach is to assess the knee reflex with the standard method and then, if a definite reflex has not been recorded, move on to use the “superior patellar supine” method; and for the ankle reflex begin with the “plantar strike method” and then, if necessary, move on to use the “Achilles strike elevated” method.