Behavioral Neuroscience1996, Vol. 110, No, 2, 300-304

Copyright 1996 by the American Psychological Association, Inc.0735-7044/96/S3.00

The NMD A Antagonist MK-801 Affects Nonspatial Learningin Preweanling Rats

David A. Highfield, Kimberly Nixon, and Abram AmselUniversity of Texas at Austin

The Af-methyl-D-aspartate (NMDA) subtype of the excitatory amino acid receptor has beenimplicated in several kinds of learning and memory, as well as in long-term potentiation (LTP), aputative cellular mechanism for learning and memory. This experiment examined the role of theNMDA receptor in patterned single-alternation (PSA) learning in preweanling rats followingintraperitoneal injections of 0.05 mg/kg MK-801, a selective NMDA antagonist. MK-801significantly inhibited PSA at both 60-s and 30-s intertrial intervals (ITIs), and attenuated, but didnot block, learning at 8-s ITI. These results are compared with effects on PSA, a form of nonspatial,memory-based learning, observed after early postnatal exposure to alcohol, infant hippocampallesions, and infant exposure to X-irradiation, and they add strongly to these earlier demonstrationsof the role of the hippocampus in learning and memory that is clearly nonspatial andnon-cognitive-map-related.

The Af-methyl-D-aspartate (NMDA) subtype of excitatoryamino acid receptor has been implicated in various aspects oflearning, from experiments examining its role in conditionedemotional response (CER; Hoehn-Saric, McLeod, & Glowa,1991), spatial learning (Butelman, 1989; Gorter & de Bruin,1992), and passive avoidance (Paoli, Spignoli, & Pepeu, 1990),to long-term potentiation (LTP), a putative cellular mecha-nism for learning and memory (Harris, Ganong, & Cotman,1984; Morris, 1989).

In this experiment we attempted to assess directly, inpreweanling rats, the role of the NMDA receptor in patternedsingle-alternation (PSA) learning, a kind of selective learningin which the rat learns to discriminate rewarded (R) fromnonrewarded (N) trials based solely on the memory (oraftereffects) of R or N on the previous trial. During thisprocedure each R-trial is followed, after a predeterminedintertrial interval (ITI), by an N-trial, and vice versa. Past workhas indicated that performance in such learning in rats ofabout preweanling age can be affected by lesions of thehippocampus (e.g., Lobaugh, Greene, Grant, Nick, & Amsel,1989), and by a number of developmental teratogens, includingpostnatal ethanol (Greene, Diaz-Granados, & Amsel, 1992)and hippocampal X-irradiation (Diaz-Granados, Greene, &Amsel, 1992,1994). The ITI at which PSA can first be shown in120 trials is age-dependent, occurring at 8 s at 11 days of ageand at 60 s in weanlings (e.g., Stanton, 1982, 1983). Thisemergence of PSA at longer and longer ITIs approximately

David A. Highfield, Kimberly Nixon, and Abram Amsel, Depart-ment of Psychology and Institute for Neuroscience, University ofTexas at Austin.

This research was supported by National Institute on Alcohol Abuseand Alcoholism (NIAAA) Grant AA07052 and NIAAA TrainingGrant AA07471. We thank Elizabeth Katz Highfield for her help withSPSS analysis.

Correspondence concerning this article should be addressed toAbram Amsel, Department of Psychology, University of Texas, Austin,Texas, 78712. Electronic mail may be sent via Internet toamsel@psyvax.psy.utexas.edu.

parallels the postnatal development of the hippocampus,corresponding in particular in the rat to the development ofthe granule cells of the dentate gyrus and their mossy fiberconnections to pyramidal cells (Altman & Das, 1965; Bayer,1980a, 1980b; Cotman, Taylor, & Lynch, 1973; Schlessinger,Cowan, & Gottlieb, 1975). This age range also corresponds,approximately, to the development of adult levels of NMDAreceptor function (McDonald, Johnston, & Young, 1990).

To assess the potential role of NMDA receptor function inPSA and its possible relation to the aforementioned effects onPSA of hippocampal lesions, postnatal ethanol, X-irradiation,and other treatments, we injected rat pups with dizocilpinemaleate (MK-801) and tested them at 3 ITIs: 8, 30, and 60 s.MK-801 is a noncompetitive NMDA receptor antagonistwhich acts as an open ion-channel blocker. MK-801 was theNMDA antagonist of choice because it passes the blood-brainbarrier when given systemically and because of the extensiveliterature indicating that, in the adult rat, it attenuates perfor-mance in several other learning tasks, including the watermaze (Gorter & de Bruin, 1992), the eight-arm radial maze(Butelman, 1989), CER (Hoehn-Saric et al., 1991), reinforcedalternation in a T-maze (Wozniak, OIney, Kettinger, Price, &Miller, 1990), and place navigation and dark avoidance (Mon-dadori, Weiskrantz, Buerki, Petschke, & Fagg, 1989). To datethere are no other studies examining the effect of MK-801 onlearning and memory in infant rats, although chronic neonatalMK-801 treatment impairs spatial learning in adult rats(Gorter & de Bruin, 1992).

Method

Subjects

Subjects were 51 Holtzman albino rat pups, 16—17 days of age at thetime of PSA learning, bred in our colony from Harlan Sprague-Dawleystock. Rats were maintained on a 14:10-hr light-dark cycle with ad libaccess to food and water. On postnatal day (PND) 3, litters were culledto 8 pups.

300

NMDA AND NONSPATIAL LEARNING 301

Behavioral Apparatus and Procedure

The procedures are similar to those of Greene et al. (1992). PSA wastested in a Plexiglas 60-cm straight runway with dark floor and sidewalls. The runway consisted of a startbox (13 x 7.5 x 12 cm), an alley(60 x 7.5 x 12 cm), and a goalbox (17 x 25 x 12 cm). All had smoothPlexiglas floors and were covered with hinged, clear Plexiglas lids. Thegoalbox was bisected by a metal gate into rear (reward, 15 x 25 x 12cm) and front (nonreward, 8.5 x 25 x 12 cm) sections. The alley wasseparated from both the startbox and goalbox by manually operatedopaque Plexiglas sliding doors. Odors were expelled by an exhaust fan(115 V, 50 Hz, 10 cm in diameter) mounted on the rear wall of thereward section of the goalbox. The alley section of the runway wasdivided into three 20-cm segments by photocell circuits allowing threemeasures of the approach response: start, run, and goal times(converted to speeds). A dark Plexiglas enclosure (12 x 12 x 18 cm)served as a holding box during the ITI.

Twenty-four hours prior to runway training (PND 15), pups wereremoved from the litter, weighed, food- and water-deprived, andplaced in individual clear Plexiglas compartments (14 x 10 x 14.5 cm)in an incubator heated to 31-33 °C. Approximately 18 hr prior to thefirst session, all subjects were fitted with oral cannulas according to theprocedure of Hall and Rosenblatt (1977). Following cannulation thepups were returned to the holding boxes.

Goalbox training, to orient the pups to the runway apparatus, wasconducted approximately 12 hr prior to the first session. This began byplacing the pups to be trained into the runway apparatus with all doorsand the metal gate between goalboxes raised, and allowing them toexplore the apparatus for 10 min. They were then confined to thecombined goalbox for an additional 10 min. Each pup was then giventhree 0.03-ml infusions of infant diet through the oral cannula, bymeans of a Harvard Apparatus (South Natick, MA, Model 906)infusion pump. An anesthetized dam was first placed in the rearportion of the goalbox. The pup was then placed on the dam's ventrumand, contingent upon attaching to a nipple, was given the first milkinfusion. For the next two infusions the pup was placed in the frontsection of the goalbox and had to crawl toward the dam and attach to anipple. Following goalbox training, pups were put back into the heatedholding boxes for the balance of the deprivation period.

Runway training was conducted in 3 groupings differentiated bynumber of trials and the ITI. It consisted of either 240 trials (threesessions of 40 trials on each of 2 days; 60-s ITI), or 120 trials (threesessions of 40 trials on 1 day; 30- or 8-s ITI). The time between sessionswas 4 hr (9 a.m., 1 p.m., and 5 p.m.). Prior to each session the pups hadtheir bladders voided and were weighed. A lactating dam wasanesthetized with sodium pentobarbital (Nembutal; 32.5 mg/kg bodyweight) and returned to the litter until the start of the session so thather nipples would be well suckled. This dam was age-matched so thatthe experimental pups were as close to the age of her actual pups aspossible. R- and N-trials were alternated such that odd-numberedtrials were R and even-numbered trials were N. On all trials the pupwas released from the startbox and allowed to traverse the alley andenter the goalbox. On breaking the last photobeam, on an R-trial, themetal gate was raised automatically, giving the pup access to theanesthetized dam. The opaque Plexiglas door between the alley andthe goalbox was then closed. When the pup attached to the nipple, theinfusion pump was activated and a reward of 0.03 ml was deliveredover 9 s. The duration of the time spent in the goalbox wasapproximately 30 s. On N-trials the aluminum gate was not raised, andthe pup was held in the front portion of the goalbox for 30 s, did notreach the dam, and received no milk infusion. Following each trial, thepup remained in the ITI box for 8, 30, or 60 s according to groupassignment.

At the end of all but the last sessions, pups were weighed andreturned to the heated holding boxes. After the third session of the

first day (60-s ITI only), the pups were given a supplemental feedingwhile attached to the anesthetized dam outside of the runwayapparatus. The amount of the supplemental diet fed was 0.6 ml(equivalent to 20 infusions); this was given in order to preventexcessive weight loss during the extended intersession interval. Afterthe final session (third session of the second day, 60-s ITI; third sessionof the first day, 8-s and 30-s ITI) cannulas were removed and pups wereplaced back in their home litters.

Four dependent measures were taken. A time measure (convertedto speed) for each of the three 20-cm alley segments (start, run, andgoal), and a total-alley speed. If a pup did not traverse a segment in 30s, it was gently pushed into the goalbox area and a time of 30 s wasrecorded for that segment and all subsequent segments on that trial.The data were collected and stored on an IBM-AT using a Profes-sional FORTRAN program.

Drug Treatment

Pups were administered 0.05 mg/kg MK-801 (Research Biochemi-cal International, Natick, MA) ip, 30 min before each session. MK-801was mixed in physiological saline and volume was adjusted to the pups'weights so that the dose given was between 0.08 and 0.12 ml of solution.This relatively small volume was used because of the small size of therat pups at this age (around 40-50 g). Each rat pup was matched to alittermate control that received only physiological saline (0.1 ml).

Analysis

Each of the four measures (start, run, goal, and total times) wasconverted to speeds (cm/s) and in every case blocked into averages offive R-trials and five N-trials. There were, therefore, 12 R-trial and 12N-trial blocks for the 8- and 30-s ITI conditions and 24 such blocks forthe 60-s ITI condition.

For the 8- and 30-s ITI conditions, each of the four measures wasanalyzed with a 2 x 2 x 12 (or 2 x 2 x 24 in the 60-s ITI condition)mixed multivariate analysis of variance (MANOVA) with rewardcondition and blocks of 5 trials as the within-subjects variables anddrug condition as the between-subjects variable. Univariate post hoctests were performed as indicated by the omnibus tests.

Results

The results are first reported as total runway speeds (a latersection will report start, run, and goal data). They are based on48 of the 51 rats. Of the 3 that were dropped, 1 died and 2 wereexposed to problems with the apparatus. In a preliminary testof MK-801, 3 pups (not included in the overall sample size)received doses of MK-801 that prevented them from traversingthe runway and attaching to the dam's teats (1 at 0.2 mg/kgand 2 at 0.1 mg/kg). This determined the dose of MK-801finally used, 0.05 mg/kg. The results are, therefore, based on 8rat pups per group. They are reported in order of increasingintertrial interval, showing that the larger the ITI (i.e., themore difficult the learning), the greater the effect of MK-801on PSA learning.

8-s ITI

Both saline- and MK-801-treated rat pups learned to dis-criminate N- from R-trials at 8-s ITI, running slower on N-than on R-trials within three 40-trial sessions, F(ll, 154) =2.75, p < .003 (see Figure 1). The saline-treated group had a

302 HIGHFIELD, NIXON, AND AMSEL

60 -

50 -

40 H

& 30^

o>•| 20-c

S 10-

0

Saline Control 8-s ITI \ N

60 -i

50-

^ 40-

1°- 31

10-Saline Control 30-s ITI

60 -

I 5°'

~ 40-•Oeu

30 H

10 -MK-801 8-s ITI

10 12

blocks of 5 trials

Figure 1. Patterned single alternation at 8-s intertrial interval: 16-day-old litter-matched pups exposed to either MK-801 (0.05 mg/kg; n = 8)or saline (n = 8). Solid squares = reward (R) trials; open squares =nonreward (N) trials; ITI = intertrial interval.

1

a

o>

60 -

50-

40-

X-

20 -

10 -MK-801 30-s ITI

10 12

blocks of 5 trials

Figure 2. Patterned single alternation at 30-s intertrial interval:16-day-old litter-matched pups exposed to either MK-801 (0.05 mg/kg;n = 8) or saline (n = 8). Solid squares = reward (R) trials; opensquares = nonreward (N) trials; ITI = intertrial interval.

final (Session 3, Block 12) average running speed of 5.17 ±3.44 cm/s on N-trials and 47.04 ± 13.10 cm/s on R-trials, F(\,15) = 13.14,/? = .002. The MK-801-treated group had a finalaverage running speed (Block 12) of 19.81 ± 8.02 cm/s onN-trials and 51.97 ± 13.25 cm/s on R-trials, F(l, 15) = 10.50,p = .005. As Figure 1 indicates, however, at an almost statisti-cally significant level, even at 8-s ITI, the MK-801 treated ratpups ran faster than saline controls on N-trials, F(4, 12) =4.24, p = .057, suggesting somewhat weaker PSA learning.

30-s ITI

At the 30-s ITI most saline-treated pups showed memory-based PSA learning on the third session (see Figure 2), with anaverage final (Block 12) running speed of 32.71 ± 21.47 cm/son N-trials and of 49.05 ± 14.36 on R-trials, F(l, 12) = 6.75,p = .020. The MK-801-treated rat pups did not, however, showpatterning after three sessions. Their final average runningspeeds on R- and N-trials were not significantly different from

each other (N = 45.35 ± 7.04, R = 50.66 ± 6.94), F(l, 12) =4.02,p = .063.

60-s ITI

Saline-injected pups trained at 60-s ITI showed PSA in thefifth and sixth sessions (see Figure 3). Their average final(Block 24) running speeds were 76.35 cm/s on R-trials and37.53 on N-trials. MK-801-injected rat pups, however, did notshow significant PSA discrimination. Although there is thesuggestion of the onset of patterning, average final (Block 24)total running speeds between N- (64.70 ± 8.67) and R-trials(73.14 ± 6.88) are not statistically different, F(l, 15) = 3.67,p = .075.

Start, Run, and Goal Times

In addition to the overall running speeds, the start, run, andgoal times (converted to speeds) were analyzed at each ITI. At

NMDA AND NONSPATIAL LEARNING 303

•O9Va.M

O)Cec3cc

80

70

60

50-

40-

30-

20-

10-Saline Control 60-s ITI

•Q4)0)aw

ccoc

80-

70-

60-

50-

40-

30 -

20-

10-

0

MK-801 60-s ITI

0 2 4 6 8 10 12 14 16 18 20 22 24

blocks of 5 trials

Figure 3. Patterned single alternation at 60-s intertrial interval:16-17-day-old litter-matched pups exposed to either saline (n = 8) orMK-801 (0.05 mg/kg; n = 8). Solid squares = reward (R) trials; opensquares = nonreward (N) trials; ITI = intertrial interval.

8-s ITI, there was a significant difference in start speeds inMK-801- versus saline-treated rat pups during the last twoblocks of N-trials, F(l, 15) = 5.02, p = .04 andf(l, 15) = 5.08,p = .04. There were no other consistent differences amongthese separate runway measures.

The differences between the MK-801-treated rats and salinecontrols in the 30-s or 60-s ITI conditions are a composite ofstart, run, and goal measures. Generally speaking, then, thetotal speed measures shown in Figures 1, 2, and 3 reflect theresults over the entire runway, not just particular segments ofthe runway. There were no consistent differences amongrunway measures between the MK-801 and control conditions.

Discussion

At all three ITIs, rat pups given MK-801 learned signifi-cantly differently from controls. In the case of the 30- and 60-sITIs, rats exposed to MK-801 exhibited no (or possibly, in the60-s case, minimal) PSA; that is, they ran as fast on N-trials ason R-trials at the stage of training in which the controls had

learned the PSA discrimination. Both the MK-801- andsaline-exposed pups learned to discriminate between R- andN-trials at 8-s ITI; however, the MK-801 rats still ran fasterthan did the controls on N-trials, suggesting slightly weakerdiscrimination. With the possible exception of start time at 8-sITI, there were also no significant differences between groupsin start, run, and goal times, suggesting that differences inrunning speed are reflective of differences throughout thelength of the runway and that all components contribute to theoverall running speed rather than MK-801's affecting just asegment of it; that is, MK-801 did not just influence latency tostart, to accelerate (run), or to decelerate (goal) in thepreweanling rat.

Like previous reports of hippocampal lesions (Lobaugh etal., 1989), X-irradiation (Diaz-Granados, Greene, & Amsel,1992, 1994), and early postnatal ethanol exposure (Diaz-Granados, Chuang, & Amsel, 1994; Wigal, Lobaugh, Wigal,Greene, & Amsel, 1988), MK-801 attenuated the performanceof rats at about preweanling age on memory-based discrimina-tion learning at 30- or 60-s ITI. It is also interesting to note thatnone of the abovementioned treatments affected PSA learningat 8-s ITI. This tends to support the conjecture that thehippocampus is not (or is minimally) engaged in this kind ofmemory-based learning at our shortest ITI and is consistentwith the intermediate-term-memory hypothesis of Rawlins(1985) and the position of Lobaugh et al. (1989).

A second possibility is that this dose of MK-801 does notinhibit all NMDA receptor activity, and that a reduced levelexcitatory amino acid transmission is sufficient for discrimina-tion at an 8- but not a 30- or 60-s ITI.

Although systemic MK-801 acts at NMDA sites throughoutthe brain, the present result, in conjunction with earlier resultsfrom our laboratory (Diaz-Granados et al., 1992; Green et al.,1992; Lobaugh et al., 1989), lends support to the propositionthat MK-801 attenuates PSA by acting at hippocampal NMDAreceptor sites. The MK-801-treated pups also showed lowlevels of ataxia at the 0.05 mg/kg dose. (They were souncoordinated at higher doses, not included in these results,that they could not attach to the dam's teats.) The ataxia maybe due to MK-801's actions at other NMDA receptor sites,including the cerebellum and cerebral cortex. To completelyrule out the effects of nonhippocampal NMDA receptors inPSA learning may involve injections of another NMDA antago-nist, such as D-2-amino-5-phosphonovalerate, directly into thehippocampal formation. However, this procedure may havecomplications in infant or preweanling rats, given that thehippocampus continues to develop during these periods,particularly the dentate gyrus, which does not reach its "adult"number of cells until about PND 30 (Altman & Bayer, 1975).

The involvement of the hippocampus in PSA argues stronglyagainst the dominant view (over almost the last two decades)that the hippocampus is primarily (or even exclusively) in-volved in cognitive-mapping functions and in spatial learning(O'Keefe & Nadel, 1978). The point is that there is no spatialcomponent in PSA learning or, generally, in learning based onany of a family of dispositional reward-schedule effects (Am-sel, 1992, 1993; Amsel & Stanton, 1980). The only clue to thenext trial in PSA learning is the "memory" of the outcome ofthe previous trial. There is no external cue. Consequently,

304 HIGHFIELD, NIXON, AND AMSEL

these results tend also to favor an interpretations of thehippocampus as involved in working memory (Olton, 1983), orintermediate-term-memory (Rawlins, 1985), rather than (or,perhaps, in addition to) cognitive mapping (O'Keefe & Nadel,1978).

References

Altman, J., & Bayer, S. (1975). Postnatal development of the hippocam-pal dentate gyrus under normal and experimental conditions. InR. L. Isaacson & K. H. Pribram (Eds.), The hippocampus: Vol. 1.Structure and development (pp. 95-119). New York: Plenum Press.

Altman, J., & Das, G. D. (1965). Autoradiographic and histologicalevidence of postnatal hippocampal neurogenesis in rats. Journal ofComparative Neurology, 124, 319-336.

Amsel, A. (1992). Frustration theory: An analysis of dispositional learningand memory. Cambridge, England: Cambridge University Press.

Amsel, A. (1993). Hippocampal function in the rat: Cognitive mappingor vicarious trial and error? Hippocampus, 3, 251-256.

Amsel, A., & Stanton, M. (1980). Ontogeny and phytogeny ofparadoxical reward effects. In J. S. Rosenblatt, R. A. Hind, C. Beer,& M. Busnel (Eds.),Advances in the study of behavior (pp. 227-274).New York: Academic Press.

Bayer, S. A. (1980a). Development of the hippocampal region in therat: I. Neurogenesis examined with 3H-thymidine autoradiography.Journal of Comparative Neurology, 190, 87-111.

Bayer, S. A. (1980b). Development of the hippocampal region in therat: II. Morphogenesis during embryonic and early postnatal life.Journal of Comparative Neurology, 190, 115-134.

Butelman, E. R. (1989). A novel NMDA antagonist, MK-801, impairsperformance in a hippocampal-dependent spatial learning task.Pharmacology, Biochemistry & Behavior, 34, 13-16.

Cotman, C. W., Taylor, D., & Lynch, G. (1973). Ultrastructuralchanges in synapses in the dentate gyrus of the rat during develop-ment. Brain Research, 63, 205-213.

Diaz-Granados, J. L., Chuang, M. S., & Amsel, A. (1994). Mitigatingeffects of combined prenatal and postnatal exposure to ethanol onhippocampally-mediated memory-based learning in the infant rat.Neuroscience Abstracts, 20, 1611.

Diaz-Granados, J. L., Greene, P. L., & Amsel, A. (1992). Memory-based learning in preweanling and adult rats after infantile X-irradiation-induced hippocampal granule-cell hypoplasia. Behav-ioral Neuroscience, 106, 940-946.

Diaz-Granados, J. L., Greene, P. L., & Amsel, A. (1994). Selectiveactivity enhancement and persistence in weanling rats after hippo-campal X-irradiation in infancy: Possible relevance for ADHD.Behavioral and Neural Biology, 61, 251-259.

Gorter, J. A., & de Bruin, J. P. C. (1992). Chronic neonatal MK-801treatment results in an impairment of spatial learning in the adultrat. Brain Research, 580, 12-17.

Greene, P. L., Diaz-Granados, J. L., & Amsel, A. (1992). Bloodethanol concentration from early postnatal exposure: Effects onmemory-based learning and hippocampal neuroanatomy in infantand adult rats. Behavioral Neuroscience, 106, 1-10.

Hall, W. G. & Rosenblatt, J. S. (1977). Development of nutritional

control of food intake in suckling rat pups. Behavioral Biology, 24,413-427.

Harris, E. W., Ganong, A. H., & Cotman, C. W. (1984). Long-termpotentiation in the hippocampus involves activation of N-methyl-D-aspartate receptors. Brain Research, 323, 132-137.

Hoehn-Saric, R., McLeod, D. R., & Glowa, J. R. (1991). The effects ofNMDA receptor blockade on the acquisition of a conditionedemotional response. Biological Psychiatry, 30, 170-176.

Lobaugh, N. J., Greene, P. L., Grant, M., Nick, T., & Amsel, A. (1989).Patterned (single) alternation in infant rats after combined orseparate lesions of hippocampus and amygdala. Behavioral Neurosci-ence, 103, 1159-1167.

McDonald, J. W., Johnston, M. V., & Young, A. B. (1990). Differentialontogenic development of three receptors comprising the NMDAreceptor/channel complex in the rat hippocampus. ExperimentalNeurology, 110, 237-247.

Mondadori, C., Weiskrantz, L., Buerki, H., Petschke, F., & Fagg, G. E.(1989). NMDA receptor antagonists can enhance or impair learningperformance in animals. Experimental Brain Research, 75, 449-456.

Morris, R. G. M. (1989). Synaptic plasticity and learning: selectiveimpairment of learning in rats and blockade of long term potentionin vivo by the N-methyl-D-aspartate receptor antagonist APS.Journal of Neuroscience, 9, 3040-3057.

O'Keefe, J., & Nadel, L. (1978). The hippocampus as cognitive map.Oxford: Oxford University Press.

Olton, D. S. (1983). Memory functions and the hippocampus. In W.Seifert (Ed.), Neurobiology of the hippocampus, (pp. 335-373). NewYork: Academic Press.

Paoli, F., Spignoli, G., & Pepeu, G. (1990). Oxiracetam and D-pyroglutamic acid antagonize a disruption of passive avoidancebehaviour induced by the Af-methyl-D-aspartate receptor antagonist2-amino-5-phosphonovalerate. Psychopharmacology, 100, 130-131.

Rawlins, J. N. P. (1985). Associations across time: The hippocampus asa temporary memory store. Behavioral Brain Science, 8, 479-496.

Schlessinger, A. R., Cowan, W. M., & Gottlieb, D. I. (1975). Anautoradiographic study of the time of origin and the pattern ofgranule cell migration in the dentate gyrus of the rat. Journal ofComparative Neurology, 159, 149-176.

Stanton, M. E. (1982). Performance of 11- and 14-day-old rats on aworking-memory problem. Behavioral and Neural Biology, 36, 304-310.

Stanton, M. E. (1983). Dissociation of patterned alternation learningand the partial reinforcement extinction effect in preweanling rats.Animal Learning & Behavior, 11, 415—423.

Wigal, T., Lobaugh, N. J., Wigal, S. B., Greene, P. L., & Amsel, A.(1988). Sparing of patterned alternation but not partial reinforce-ment extinction effect after prenatal chronic exposure to ethanol ininfant ruts. Behavioral Neuroscience, 102, 43-50.

Wozniak, D. F., Olney, J. W., Kettinger, L., Ill, Price, M., & Miller,J. P. (1990). Behavioral effects of MK-801 in the rat. Psychopharma-cology, 101, 47-56.

Received May 22,1995Revision received July 14,1995

Accepted July 17,1995 •