Effects of Partial Reinforcement of a Discriminative Response on Extinction in the Turtle:Spaced-Trial Acquisition Training

Two groups of 10 turtles(Geoclemys reevesii)each were trained in a black-white discrimination situation using a Y-maze,one with 50% partial(PR)and the other with continuous reinforcement(CR),followed by extinction of the learned response.They received widely spaced trials(one trial per day)during acquisition training and massed trials during extinction.The CR group learned the task more rapidly than the PR group is acquisition but both groups resulted in almost indentical asymptotic level.Significant differences in extinction were found in terms of both of choice and running latency measures.Extinction curves for the two groups demonstrated rather typical partial reinforcement extinction effect(PREE).But the difference in choice measure between groups can be interpreted as the difference in running latency because of our definition of incorrect choice.The results were compared with those obtained in analogous experiments with turtles and discussed in terms of the trial spacing during extinction used in the present study and its relation to the assessment of a spaced-trial PREE.|1群10匹の2群からなるクサガメをY型迷路を用いて白黒弁別事態のもとで訓練し,そののち消去試行を与えた。1群は50%の部分強化(PR),他の1群は連続強化(CR)である。2群とも習得訓練で1日1試行の分散試行が,消去試行では集中試行が与えられた。習得訓練ではCR群がPR群よりも学習が速かったが,最終的には両群ともほぼ同一の漸近値に到達した。消去では選択数及び反応潜時の両測度において有意差がみられた。消去における両群の遂行曲線は典型的な部分強化消去効果(PREE)の生起をあらわしているといえよう。ただし本実験が用いた誤反応に対する手続のゆえに,正反応選択数の測度における上述の差はむしろ走行潜時の差を反映するものと解釈される。これらの諸結果はカメを使った他の実験結果と比較検討され,その上で本実験で用いた消去時の試行分散の条件と,それの分散試行PREEの生起判定への関連とが考察された。