Ifthisorderbenecessary，wheneverasubjectexperiencestheformeroftwomovementsthusconnected，itwill（invariably）
experiencethelatter;if，however，theorderbenotnecessary，butcustomary，onlyinthemajorityofcaseswillthesubjectexperiencethelatterofthetwomovements。Butitisafactthattherearesomemovements，byasingleexperienceofwhichpersonstaketheimpressofcustommoredeeplythantheydobyexperiencingothersmanytimes;henceuponseeingsomethingsbutoncewerememberthembetterthanotherswhichwemayhavebeenfrequently。
Whenevertherefore，wearerecollecting，weareexperiencingcertainoftheantecedentmovementsuntilfinallyweexperiencetheoneafterwhichcustomarilycomesthatwhichweseek。Thisexplainswhywehuntuptheseries（ofkineseis）havingstartedinthoughteitherfromapresentintuitionorsomeother，andfromsomethingeithersimilar，orcontrary，towhatweseek，orelsefromthatwhichiscontiguouswithit。Suchistheempiricalgroundoftheprocessofrecollection;forthemnemonicmovementsinvolvedinthesestarting—pointsareinsomecasesidentical，inothers，again，simultaneous，withthoseoftheideaweseek，whileinotherstheycompriseaportionofthem，sothattheremnantwhichoneexperiencedafterthatportion（andwhichstillrequirestobeexcitedinmemory）iscomparativelysmall。
Thus，then，itisthatpersonsseektorecollect，andthus，too，itisthattheyrecollectevenwithouttheeffortofseekingtodoso，viz。whenthemovementimpliedinrecollectionhassupervenedonsomeotherwhichisitscondition。For，asarule，itiswhenantecedentmovementsoftheclassesheredescribedhavefirstbeenexcited，thattheparticularmovementimpliedinrecollectionfollows。
Weneednotexamineaseriesofwhichthebeginningandendliefarapart，inordertoseehow（byrecollection）weremember;oneinwhichtheylienearoneanotherwillserveequallywell。Foritisclearthatthemethodisineachcasethesame，thatis，onehuntsuptheobjectiveseries，withoutanyprevioussearchorpreviousrecollection。For（thereis，besidesthenaturalorder，viz。theorderofthepralmata，oreventsoftheprimaryexperience，alsoacustomaryorder，and）bytheeffectofcustomthemnemonicmovementstendtosucceedoneanotherinacertainorder。Accordingly，therefore，whenonewishestorecollect，thisiswhathewilldo:hewilltrytoobtainabeginningofmovementwhosesequelshallbethemovementwhichhedesirestoreawaken。Thisexplainswhyattemptsatrecollectionsucceedsoonestandbestwhentheystartfromabeginning（ofsomeobjectiveseries）。For，inorderofsuccession，themnemonicmovementsaretooneanotherastheobjectivefacts（fromwhichtheyarederived）。Accordingly，thingsarrangedinafixedorder，likethesuccessivedemonstrationsingeometry，areeasytoremember（orrecollect）whilebadlyarrangedsubjectsarerememberedwithdifficulty。