sign in sign up
<<
Home , Big Clifty Formation, Biospeleology, Burrow, Cave survey, Grassy Cove, Grotto

GEOLOGICAL SURVEY UNIVERSITY OF KENTUCKY, LEXINGTON Donald C. Haney, State Geologist and Director

CAVES and

KENTUCKY

SPECIAL PUBLICATION 12 Series XI, 1985 ISSN 0075-5613

KENTUCKY GEOLOGICAL SURVEY UNIVERSITY OF KENTUCKY, LEXINGTON Donald C. Haney, State Geologist and Director

CAVES AND KARST OF KENTUCKY

Percy H. Dougherty, Editor

Published in cooperation with the National Speleological Society

COVER PHOTOGRAPH

Creek section of Sloans . Pulaski County, Kentucky. (Photo by K. L. Day$

SPECIAL PUBLICATION 12 Series XI, 1985 John G. Beard, Geolog~stIV, Henderson Offlce Wayne 1. FranKle, Geologlst I1 Art Gal laher, Jr., Chancellor, Lexlngton Campus Patr~cKJ. Bood~ng, Geologlst I1 Wlmberly C. Koyster, Vlce Chancellor Kesearch JacK R. , Geolog~st11 and Dean ot the Uraduate School Brandon L. Nuttall. Geolog~stI1 James Y. McDonald, Execut~veD~rector, Unfverslty Jul~eR. Kemper, Geolog~stI of KentucKy Research Foundat I on Frances Benson, L~brarylechn~clan 111 Robert H. Dan~el,Laboratory lechn~c~anB KENTUCKY GEOLUG ICAL SCIWEY Jacquel~neH. tmbry, Data Entry Uperator 111 AVV ISORY BOARD VICKI F. Campbell, Dratt~nglechn~ctan Ph~lM. M~les,Chairman, Lexlngton Jane ball lon, JenKfns Water Hesources beCtlOn: Wallace W. Hagan, Lexlngton James S. D~nger,Head Henry L. HlnKle, Par15 James K~pp,Geolog~st 11 B. W. McDonald, Pafntsv~lle H~chardS. Smalles, Ueologlst I1 W. A. Mossbarger, Lextngton Margaret A. Townsend, tieolog~stII W~lllamJ. Reynolds, fillen Henry H. Spald~ng.Hazard Com~uterServfces Group: Henry D. Stratton, P~Kevllle Steven Cordlv~ola.Geoloq~st I11 Ralph N. Ihomas, Uwensboro Joseph B. DI xon , systems-programmer George H. , Jr., Owenshoro Elmer Wh~taKer, Lexlngton SPECIAL PHOJEC-IS DlVISILW4 Projects KEN rUCKY GEULUCil CAL 8LlWE.Y Env tronmental Protect ion Agency--Devel opment of a Donald C. Haney, D~rectorand State Ueologlst Comprehens~veUil and Gas Injection John D. Kiefer, Asslstant State Oeologlst Inventory, KentucKy James S. D~nger,Pr~nctpal Investlgator ADMINISTRA'rIVt DIVISION FranK H. Walker, Co-Prtnc lpal Investigator Personnel and Finance Sect~on: James L. Hamilton, Adminlstrattve Staff Offlcer 11 U.S. Env~ronmentalProtect~on Agency--Area of Margaret A. Fernandez, Account Clerk 'J Rev1 ew and lnjec t ton Pressure Rssessment ot Ui l and Gas Inject on Wel Is, KentucKy Clerical Sectlon: James S. O~nger,Pr~nctpal lnvest~gator Oosha B. Boyd, Staff Assistant VI H~chardS. Smalley, tieologlst 11 Donna C. Ramseur, Statt Asststant V1 bh~rleyU. Black, btatt Ass~stantV 1J.S. Geological Survey--6sslstance ~n tiather~ng Jean Kelly, Statt Hss~stantV Oata on Kentucky Coal Resources for the Juan~taG. Smlth, Staft Asslstant V, Henderson National Coal Resources Oata System Utf Ice Russell fi. Brant, Prlnclpal Investlgator

Publ lcations Sectlon: U.S. Geol og~cal Survey--Coal Sampl ing ~n the Donald W. Hutcheson, Head Western KentucKr Coal Field Margaret K. Luther, Ass~stantEditor James C. Cobb, Pr~nc~palInuest~gator Roger B Potts, Chtef Cartograph~cIllustrator James C. Currens, Co-Pr~ncipal Investigator Robert C. Holladay, Drafting Technlclan Will lam A. Brlscoe, Ill, Sales Supervisor KentucKy Department ot MI l I tary Atfa~rs--Drainage Roger S. BanK5, Account Clerk 11 Determ~nat~ontor the Boone Nat~onalGuard John Dav~s,Stores WorKer Center, 6ranKltn County, Kentucky PatrlcK H. McHaftie, Geolog~st/Feographer I1 James S. D~nger,Prlnc~pal Investlgator J. V. Ihra~lK~ll,Lo-Prlnclpal Investlgator GEOLOGlWL DlVISIUN Coal Sect~on: KentucKy t'latural Resources and Env~ronmental James C. Cobb, Head Protect~onCablnet--Vel lneat lon and Russell A. firant, Geolog~stV Documentat~onot M~n~ng-RelatedSubsidence In Allen L). W~lllamson, Geologfst 10, Henderson HopK~ns, Ohlo, Un~on,and Webster Uountles Otf I ce Hlchard E. Sergeant, Pr~nc~palInvest gato or Donald H. Chesnut, Jr., Geologlst 111 Hlchard A. Smath, beologlst I James C. Currens, Geolog~st111 John 6. SttcKney, Ueologlst 1 Hlchard t. Sergeant, Geologfst 111 Apr~l L. Cowan, Field Asslstant Davfd A. W~lllams,tieologlst 111, Henderson Ofttce Hlchard A. Smath, Geologtst I Gas Research Inst~tute--Study of Hydrocarbon John F. St~cKnev,tieologlst 1 Productlon from the Devonlan Shale fn Letcher, Aprfl L. Cowan, Geology F~eldAsslstsnt Knott, Floyd, and PlKe Counttes, Eastern KentucKy lndustr~aland Metal l lc M~neralsSectton: Wayne 1. FranKce, Pr~nc~pallnvestfgator Garland R. Vever, Jr., Head JacK H. Moody, Geologlst I1 Eugene J. , Geolog~stIV Julle R. Kemper, Geolog~stI Warren H. Anderson, Geolog~st11 Jacqueline H. kmbry, Data Entry Operator I11 'JICKI k. Campbell, Uraftlng Techn~c~an Strat~araphyand Petroleum Geoloqy Sect~on: John D. Kieter,. Acttng Head and Asslstant State L1.S. Geol oglcal Survey--Mldcont lnent Strategic and Geolog~st Crl t lcal Mtnerals Program Martln C. Noger, Geologlst V Warren H. finderson, Prlnclpal Inuestlgator FranK H. WalKer, Ueologlst IV Garland K. Dever, Co-Pr~nc~palInvestfgator CONTENTS Page

Preface ...... 1 AcKnowledgments ...... 2 Foreword...... 3

Section f Chap t~r1 : tiverv i ew o+ the 6eol ogy and Physl cal of Kentucky ...... 5 Chapter 2: Caves o+ KentucKy ...... 18

Section 11

Chapter 3: The Inner Blue Grass Karst Heglon ...... 28 Lhapter 4: Patterns ot Cavern Development Along the Cumberland Escarpment ln Southeastern KentucKy ...... 63 Chapter 5: Caves ot Northeastern Kentucky (W~thSpecial Emphas~son Carter Caves State Park) ...... 78 Chapter a: Ptne Mounta~nKarst and Caves ...... 86 Chapter 7: The Mammoth Cave Reg1 on and ...... 97 Chapter 8: Western Kentucky Heglon ...... 119

Section 111

Chapter 9: Cave Ll f e ot Ken tucKr ...... 146 Chapter 10: Vertebrate Remains tn Kentuckr Caves ...... 16.9 Chapter 11: Archeologv ...... 176 Chapter 12: Caves and the Saltpeter Industry in KentucKy ...... 187 PREFACE Members o?: the National Speleological Society are to be commended for the preparat ion o+ this pub1 i cat ion on the caves and related Karst features of KentucKy. Percy H. Oougherty, who was respons~ble for ini t iating the project, dolng technical edit~ngof the manuscripts, and Surnishing typ,escr i p t wh i ch was used as pri n ter's copy, deserlJes spec I a1 recogn i t I on.

The role o+ the KentucKy Geological Survey was to do rnlnor copy edl t I ng, prepare the I ayuut, paste up camera- ready copy, and maKe arrangements for getting the manuscript printed. ln order to meet the printlng dead1 ~ne,it was necessary tc~utilize computer pr~ntout,prepared by the Nat i ona'l Spe 1201 ugl cal Soc 1 ety, as typescr I p t +or the pub1 i cat I on. Therefore, the pub1 i cat i on does not conform to the usual sty1 e and edr tor I al standards of the KentucKy Geol ogi cal Survey.

It is telt that thls publ~cat~onw~ll sat15.f~ a wl despread need +or up-to-date I nformat I on about caves and Karst geology ~n KentucKy. The d~verserange of top I cs covered should be o+ Interest to a wlde aud~ence, ~nclud~ng both sclent lsts and laymen.

Dona1 d W. HIJ tcheson KentucKy Geol ogl cal Suruey Lexington, KentucKy June 19e5 ACKNOWLEDGMENTS Without the help of many people, this booK would not have been possible. The authors of the chapters should be congratulated for worKing under a tight schedule and nearly impossible deadl ines. In addi t ion, the edi tor is deeply indebted to the f ol 1owing col l eagues who cri t i cal l y reviewed sections of the booK and made many excellent suggestions: Steve Justham, Ron Lliiamarter, John Ho+felt, Patrick Munson, A1 Scheide, George Crowthers, Horton Hobbs, 111, Don Pol locK, Harold Meloy, George Moore, Fred Grady, Russell Graham, Joe Saunders, George Hupper t , Angel o George, and Charles Bishop. Donna Moore, secretary o+ the Geography Department at Kutztown University, deserves much credit +or spending many hours typi ng the text. The author also wishes to thank the Geography Department and admrnistration at Kutztown Universi ty for providing the fac i l i ties and stimulating environment in which to complete this project. The KentucKr Geol ogi car Survey a1 so deserves 5.pec i a1 recogn i t I on. Dona1d C. Haney, Di rector and State Geol ogi st, and Donald W. Hutcheson, Editor of the Survey, have provided guidance for this worK since i ts ~nceptlon. In addition, several staff members have prou I ded uluabl e assi stance i n editing the text and laying out the graphics. Finally, 1 am indebted to my wife, Anne, and sons, Thomas and Hobert, +or the I r continuing support and undPrstandinq. This bcnoK 1s dedicated to them and my father, Percy H. Dougherty, Sr., who is sadlr missed since his death in Aprr I.

Percy H. Dougherty, edi tor

Kutzto!,~~nUn~versi ty

Jl~ne1985 FOREWORD

CAVES AND K6HST OF KENTUCKY is in applied areas; including pale- a special publication of the Na- ontology, archeology, history, and tional Spel eologi cal Society and biology, (6) to present extensive the Kentucky Geological Survey to bibliographic material, and, (7) commemorate the Annual Convention to discuss gaps in the literature, of the National Speleological So- thereby possibly stimulating fur- ciety at Kentucky State University ther research in Kentucky cave and in Frankfort, Kentucky on June 22- karst environments. 29, 1985. It is appropriate to These goals played a major role hold the Convention in Kentucky in structuring the book into three since it is the home of Mammoth parts. Section I introduces two Cave, the world's longest cave, chapters of background information and the site of many other large necessary to understand the caves caves and karst landforms. In ad- and karst of Kentucky. Chapter 1, dition, Kentucky has a long tradi- by Percy H. Dougherty, investi- tion of cave and karst research. gates the geology and geomorphol- The Convention and this book are ogy of the State in order to give dedicated to the explorers, map- the reader a basic background in pers, and researchers who have order to understand the where and added to our knowledge and appre- why of cave formation. The discus- ciation of the State's caves. sion takes the reader back 400 When plans were being developed million years and traces the de- for the 1985 Convention, it was velopment of the present rock realized that there was a need for types and landforms so one can a book on the caves and karst of appreciate the processes that have Kentucky. After reviewing the pro- created Kentucky's great caves. f essional literature and examining Angelo Gearge expands upon this publications, a set of background by discussing where the goals was formulated to guide the known caves are located and analy- design of the book. The goals in- zing the potential for further cluded: (1) to provide a state of di scover ies . the art approach to what has been Section I1 divides the State done in Kentucky cave and karst into several cave and karst reserch, written by people who are regions. Each region is discussed doing the work; (2) to accumulate by an individual who has done sub- diverse materials about Kentucky stantial research in the area. The caves and karst in one publica- Elue Grass Region is written by tion, making available a quick and John Thrailkill who has researched easy reference and research vol- the karst of the ume, (3) to fill a gap in the pro- Lee ington area. The Cumber 1 and fessional literature for no single Plateau is divided into two dis- statewide reference to Kentucky tricts; a northern area centered caves and karst exists, although on Carter Caves, presented by John the Kentucky Geological Survey has Tierney; and a southern region, published several good case stu- with its major cave area around dies on caves; (4) to use the Pulaski and Rockcastl e counties regional approach to compare and described by Ralph Ewers. Although contrast various cave and karst part of the same geologic and geo- regions in Kentucky, enabling the morphic regions, the two areas are reader to appreciate karst proces- discussed separately because of ses and understand how regional differences in their differences create unique land- and their geographic separation scapes; (5) to show the status of into different drainage basins. Kentucky cave and karst research The chapters on the FOREWORD

Plateau are divided in a similar Thomas Earr, shows the status of manner. Arthur Palmer presents an research in cave biology. The in-depth chapter on the Mammoth fragile ecosystem is explored and Cave area and adjacent parts of the reader is made aware of the the Pennyroyal Plateau, while John unusual organisms inhabiting Mylroie and Mike Dyas discuss the Kentucky caves today. caves and karst of the Land Caves and karst features form Between the and the westward unique environments which are extension of the Mammoth Cave and easily disturbed. Cave formations Pennyroyal plateaus. Although once broken and removed from the similar geological 1y, the result- cave may never regenerate, result- ing caves differ substantially ing in the loss of a beautiful because of the factors producing creation of nature for future them. Local hydrology and subtle generations. Such destruction may differences in the geology between also eliminate research objects areas may result in a much differ- essential in unravelling the ent end product. Another chapter mysteries of cave processes. In investigates the cave and karst addition, undue traffic in caves processes operati ng on Pine may have a negative effect upon Mountain. Joseph Saunders has done bat colonies, endangering a much and investigation of valuable which does more thi5 unique thrust faulted region good than harm through its eradi- and shares his experiences in cation of harmful . Other Chapter 6. cave organisms may also be ser- iously reduced or wiped out by The subject matter in Section careless action. Chemical spi 11s, I11 is diverse. This section sewage, over-fertilization of farm focuses upon the applied research fields, and many other human ac- that has taken place on the caves tivities may have a severe impact and karst of Kentucky. Ron Wilson on the we1 1-being of caves and looks at the early life in Ken- karst, so we must be extra careful tucky by investigating the cave and protect these areas. paleontological record. He ex- The National Spel eologi cal So- plores how vertebrate bones got ciety is the leading organization into caves, what the presence of engaged in the conservation of the bones indicates, what animal caves and karst areas. Their slo- species once lived in Kentucky, gan is, "Take nothing but pic- and concludes with a general over- tures, leave nothing but foot- view of the paleontological work prints, and kill nothing but done in the State. Patty Jo Watson time." They have embarked upon an presents material on the early educational campai gn to develop an human population of the State, appreciation among lay people of concentrating on the archeology of the delicate nature of caves. The the Mammoth Cave Region, with Society sponsors several conferen- reference to other parts of the ces and workshops each year and State. She explores why primitive publishes two journals, the NSS people were interested in the NEWS and the NSS BULLETIN. For caves, shows evidence of their further information on how you can exploration in Mammoth Cave, and help in the conservation of caves discusses how cavers can help and karst environments, please archeological research. Stanley contact: The National Speleolog- Sides discusses the saltpeter ical Society, Cave Avenue, industry of Kentucky. Although Huntsville, 35810. concentrating on the Mammoth Cave area, the paper also discusses the impact of the saltpeter industry on the development of the United Percy H. Dougherty, ed. States and the early economy of Department of Geography Eentucky. The final selection, by Kutztown University, Pennsylvania Chapter 1 AN OVERVIEW OF THE GEOLOGY AND OFKENTUCKY Percy H. Dougherty Department of Geography Kutztown University Kutztown, Pennsylvania 19530

Kentucky is the home of the big the area in order to understand caves, a fact that cannot be where and why caves form. In an doubted when one realizes that area with flat 1ying rocks of Mammoth Cave, at nearly 500 km is homogeneous compositon, there 300 km longer than its nearest would be little need for such a competitor. Of the 10 longest study because the existance of caves in the , 2 are caves would be limited. Kentucky, found in Kentucky; this number on the other hand has a great would have been 3 if it had not diversity of earth materials and been for the recent connection of an ideal landform assemblage for the 70-km-long Hoppel System with the formation of caves and karst. Mammoth Cave. Feople working on Several factors are investigated the Mammoth Cave Project feel that throughout the book to show why the eventual mapped distance of Kentucky is the home of the "big the system wi11 exceed 900 km by ones." These factors include: (1) the turn of the century. Many composi ton of the 1i mestones-- other large cave systems in the porosity, permeability, thickness, State exceed 3 km, the distance impurities, jointing, and bedding, needed to have a cave listed on (2) composition of the caprock and the World's Longest Cave List adjacent strata, (3) the geologic published by the International structure and contribution to the Speleological Union. In fact, enhancement of speleogenetic pro- Kentucky has 45 caves on the cesses, (4) other factors such as list. weather, climate, and the impact Why has Kentucky been so of vegetation on karst processes, blessed with such long caves? and (5) differential and Why are there so many caves in the impact of river systems on Kentucky? Where are the caves karst processes. located? How have they formed? Formation of caves and karst And, what is the importance of features is a complicated process caves and their associated land- involving the interaction of many scape? These and other questions factors. In the following will be answered during the course discussi on, on1 y those caves of his chapter and examined in developed in 1i mestone more detail throughout the book. environments will be considered. Chapter 1 sets the stage on Most caves in Kentucky are the which all cave forming processes result of karst processes, work. One must know the general although there are examples of geology and geomorphic history of caves developed by tecton'ic GEOLOGY movement, piping of sediments, and deposits that form the layers of basal . In the rock we see today. The story of context of this book, a strict their deposition is told by the definition of karst will limit the many plant and animal discussion to those environments found in the rock, indicating that in which results in during the Period deep the solution of rock through seas once covered Kentucky. Later carbonation. Karst landscapes are deposi ts of the Pennsy 1vani an, often characterized by extensive Cretaceous, and Tertiary periods development of , interior show that water levels fluctuated drainage, lack of surface , greatly from near shore deposits caverns, solution sculptured rock to coastal plain, deltaic, and (karren), large springs, and other beach deposits. Great coral reefs landforms assocated with such of the Middle and Middle areas. For a more detailed periods resulted in discussion of caves and karst in several deposits in the general, the reader is encouraged State. Of the material deposited to refer to one of the following: in the ancient seas, it is the Jennings (1971), Sweeting (1972) , limestone in which we are Jakucs (1977), Sullivan and Moore primarily interested, because most (19781, or Bogli (1980). caves and karst in the The remainder of this chapter State are formed in this rock will discuss the spatial type. By understanding the spatial distribution of rock types and distribution of limestone of landforms in Kentucky. An different ages, and studying its understanding of the geologic relation to the surface, one can history and the depositional better understand where caves are enables one to to be found and how they are better understand why some areas formed. have more caves. In addition, the explanation of differential The oldest rocks in the State erosion of the various rock types of Kentucky are those exposed in and a discussion of regional the Inner Blue Grass Region around structural variations help one to Lexington. They date back over 400 understand why certain land+orms millionyears to the Ordovician are more conducive to cave and Period. In order to give a point karst formation. Geomorphic of reference, Figure f shows the history is discussed in relation geologic calendar that will be to geology in order to facilitate referred to when the age of the separation of the State into various rocks and events are distinct regions. The potential mentioned. The areal extent of the for caves and karst in each region Ordovician rock, shown in the is introduced in this chapter and geologic map (Fig. 2), covers most more fully discussed in Section I1 of the area between Lexington, for those areas containing caves Louisvi 11 e and Cinci nnat i . The and karst. rocks of the Ordovician Period were deposited in deep seas. GEOLOGY Toward the end of the period, seas became shallower, as evidenced by Virtually all of Kentucky is the amount of mud that was composed of depasited and subsequently +armed from materials that were hardened into shale. The f 011 owi ng originally deposited as sediments Si1 urian Period was characterized in great inland seas or as by warm, clear, shallow seas, as near-shore deposits in ancient indicated by the profusion of oceans. These sediments have been coral deposits and brachiopods in compressed and cemented into the Si 1ur i an do1 omi tes and 1i mestones, variety of , , although the presence of shale conglomerates, shales, and coal beds suggests periods of turbid OVERVIEW OF THE GEOLOGY AND PHYSICAL GEOGRAPHY OF KENTUCKY 7

surface running from Cincinnati, Ohio, to Nashille, . The Cincinnati Arch cuts across Kentucky and divides it in half, resulting in the +ormation of independent eastern and western basins. A north-south division between a higher zone near Lexington, Kentucky and a higher zone near Nashvi 11 e, Tennessee, also exists. This lower gap or "saddle" along the crest of the Cincinnati Arch in south-central Kentucky, results in the preservation of Mississippi an material that was stripped from the higher areas by erosion. Warping of the Cincinnati Arch continued into the Devonian Period, during which more coral rich sediments typical of a Figure 1. Geologic time chart shallow sea were laid down. Before showing the ages of rocks ex- the end of the period the sea posed in Kentucky (from McGrain, f 1oor was covered with an organic 1983, p. 3). black muck that formed a pronounced shale layer, a distinctive characteristic of the water must have occurred. Also Devonian. The black shale that during this period there was outcrops in a zone around the Blue widespread warping of the strata, Grass Region of Kentucky may have which resulted in the up1if t of important economic implications, the Cincinnati Arch, a long, for it is an oil shale (McGrain, gentle arching of the earth's 1983).

Alluv~urn(narrow strlps not shown) ["," Tertiary Cretaceous 0M~ss~ss~pp~an Devontan / S~lurlan

c;;s

n 50 100 MILES

Figure 2. Geologic map showing the age of rock outcrops in Kentucky (from McFarlan, 1958, p . 5) . GEOLOGY

As the Missi ssi ppi an Per i od geologic record of the late progressed, the deposition of the . By the end of the black shales was replaced by an Paleozoic, the Cincinnati Arch had influx of sands, gravels, silts, formed, the western and eastern and muds deposited by rivers coal basins were present, the Pine eroding the nearby land masses. Mountain Fault had been thrust 19 Deltaic deposits, with their to 16 km to the northwest, and characteristic crossbedding and many of the high-angle faults of water current worked materials central and western Kentucky were were laid down. Long periods of present. Most of the State's clear, calm conditions f 01 lowed; depositional and diastrophic and massive beds of limestone, in activity came to a close. which many of today's caves are The only major deposition located, were deposited. A period following the Paleozoic has been of recession of the seas occurred, the Gulf Embayment flooding, f 01 lowed by extensive erosion o.f during the Cretaceous Period, of the land surface. These events are the in extreme recorded as an uncomf ormi ty western Kentucky. Since that time, between the Mississippian and only alluviation from the major later Pennsylvanian strata. streams, aeol i an deposit i on of Pennsylvanian conditions were 1oess, and minor Flei stocene warm and contained intermittent deposits in have transgressions of the sea, as occurred. None of these deposits evidenced by the presence of have an impact on cave and karst marine deposits in the formation or their spatial predominantly fresh-water distribution. The major process at sediments. The period was work on Kentucky's landforms has characterized by the formation of been differential erosion, the great swamps and forests, as shown erosion of rock of different by the prolific record. resistance at varying rates. Great masses of vegetation were Conglomerates and standstone, buried under deltaic deposits and because of their greater silts. This vegetation, in the resistence, have emerged as the absence of , changed into higher landforms, such as cuestas, coal. Similar developments escarpments, and mountains; shales occurred in both the eastern and and limestones, because of their western basins on either side of weakness and susceptibility to the Cincinnati Arch, resulting in solution, form the valleys and the two major coal fields, which lowland plains. Since most of the have made the State a leading State receives in excess of 965 mm producer of coal. of precipitation a year, with some areas experiencing over 1,270 mm a The Permian Period may have year, the process of erosion has been well represented in Kentucky been rapid and has left a great but its record was undoubtedly imprint on the present landforms. erod~daway, for only minor Kentucky's modern landforms are evidence of its presence remains. therefore the remnant of past The rock of this period is landforms rather than landscapes pr'eserved in some small fault that were uplifted to their blocks and in small igneous dikes present position. These stages are in Elliott County in eastern summarized in Figure 3 which shows Kentucky and Caldwell and the regional evolution of the Crittenden counties in western Kentucky landscape. Kentucky; the only non-sedimentary rock found in the State (McGrain, A major factor in the present 1983). At the end of the period a landscape formation has been the series of uplifts occurred, river patterns and their leading to increased erosion, subsequent entrenchment. Most which obliterated much of the rivers in Kentucky do not flow OVERVIEW OF THE GEOLOGY AND PHYSICAL GEOGRAPHY OF KENTUCKY 9

FIG 2 CLOSE OF PALEOZOIC

I<. LC"lL

EARLY TERTIARY

.OTT>~,LLI 1..101~0*( . c~~~~o~~~~~e RECENT - LIIII, ..*0110*I

Figure 3. Regional evolution of the Kentucky landscape (from McFarlan, 1943, p. 159). along the dip of the sedimentary Hamilton, Ohio. Both rivers then strata. Major rivers such as the flowed north through Dayton to Kentucky, Green, Cumberland, and enter the Teays drainage from West Licking flow across the strata as superposed streams that pre-date the present landforms. The rivers were able to maintain their courses over the more resistant strata being uncovered by differential erosion. The most recent event of importance to the cave forming process was the development of the Ohio River during the . Prior to the Pleistocene, the Licking and Kentucky Rivers were part of the Teays River System. The Kentucky River flowed along the present course of the Ohio River to the area near fd RG N A - f Lawrenceburg, Indiana, where it ------___ - - - k----- flowed through the present valley of the Great Miami River. The Figure 4. Map of the preglacial Licking flowed through Cincinnati, drainage of the Teays-age basin via the present Mill Creek Valley, in Ohio and the midwest (from and joined the Kentucky River near Thornburg, 1969) . REGIONAL DIVISIONS

Virginia and Eastern Ohio at a Ordovician outcrops that form the junction in west-central Ohio. The axial portion of the Cincinnati then flowed across central Arch. Here one finds the oldest Indiana as the River Teays. surface rocks in the State, Glacial advance blocked the primari 1y 1imestone and shale. northward flow and the present Near the edges of the Blue Grass Ohio River breached the Madison some areas of Si lurian and Middle Gap to flow in its present course. Devonian limestone and shale are Because of the great flow of water included. The resulting landscape in the new channel, downward presents a gently rolling surface cutting was rapid. This cutting that appears re1ati vel y f 1at in left branch tributaries of the comparison to surrounding ancestral Ohio discordant with the thoroughly dissected regions. main stream: rapid cutting This area has been referred to as occurred resulting in the the Lexington Peneplain in the entrenchment of the Kentucky, past because of its lack of Cumberland, Tennessee, Green, and substantial re1ief (McFarlan, other rivers. Fluctuations in 1943). Regions of greater re1ief river flow and periods of rapid develop in the outcrops of shale entrenchment 1owered base 1eve1 that surround the limestone throughout the major cave areas of interior of the region. The Kentucky, a process that has Lexington and Cynthiana limestones resulted in the prof usion of occupy the central position around levels in Kenturky caves. the City of Lexington and constitute a fertile limestone REGIONAL DIVISIONS basin called the Inner Blue Grass. This basin is surrounded by a zone The end result of all od the of shales called the Eden Shale periods of deposition, uplift, Belt, which, in turn, is bounded faulting, and erosion is seen in by the Outer Blue Grass, a region the landform model in Figure 5. A of limestone and shale that wide variety of plains, plateaus, extends to the Ohio River in the escarpments, and mountains have north where it is mantled by a been sculptured out of the thin veneer of Pleistocene glaci a1 sedimentary rocks of Kentucky. deposits with 1i ttle topographic Note the similarity between the significance to the landforms of geology map (Fig. 2) and the the area. Where limestone is the distri buti on of present 1and+ orms. dominant rock, the surface is Each of these regions will be mildly karstic, and small caves discussed in detail in the may be found. The greatest relief following sections, with is not caused by karst landforms, particular attention given to but by the deeply entrenched their potential for cave and karst courses of the Kentucky, Licking, formation. The Lexington Plain or and Ohio rivers. Their courses are Inner Elue Grass, the Cumberland superimposed up to 150 meters Plateau, Pennyroyal Plateau, the below the Blue Grass surface. The Mammoth Cave Plateau, and the Pine canyons display beautiful examples Mountain Region offer the best of entrenched meanders and, in the conditions for cavern formation zone of shales, deep gorges have because of their rock type, formed. structure, and hydrologic The northwestern portion of the conditions. reqion surroundinq Loui svi 1le is not geologically kimilar to the Flue Grass but is included because Blue Grass Region ~f its topographic similarity and occurrence of the same The Blue Grass is the central geographical factors responsible lowland of Kentucky, composed of for the historical development and OVERVIEW OF THE GEOLOGY AND PHYSICAL GEOGRAPHY OF KENTUCKY 11

WESTERN KENTUCKY I EASTERN KENTUCKY MISSISSIPPI EMBA YMEN

Figure 5. Physiographic map of Kentucky showing geomorphic regions and major escarpments. Generalized cross section depicts the structure of rock units from west to east across the State. (From EIcGrain, 1983, p. 13.) economy of the Blue Grass. The brown silt loams with a high Louisville part of the Blue Grass, phosphatic level. The Woodburn called the Scottsbuvg Lowland, has Formation which once supported developed on 1imestones of Middle commercial production of phosphate Si lurian and Middle Devonian age rock. These formations have and more recent alluvium deposited resulted in the Blue Grass by ancestral stages of the Ohio becoming wide1 y known as an River. Geographically, the region agricultural island in central constitutes one of the anchors of Kentucky, characterized by its what Raitz (1980) has called the beautiful mansions and extensive "urban triangle" of the Blue horse farms. In addition, the Blue Grass, a joining by transportation Grass is the center of the burley arteries of Loui svi 11e, Lex ington tobacco region. Kentucky produces and Covington, the three largest 80 percent of all burley tobacco urban areas in the State, and all grown in the United States, making within the Flue Grass. tobacco Kentucky's number one Limestone within the Blue Grass legal agricultural crop. This crop is not only responsible for the accounts for 85 to 90 percent of karst features and caves, it has the state's agricultural income also played an important role in per year and 20 percent of the the economic growth of the region. total United States tobacco formed on the Inner Blue production in a good year (Raitz, Grass are exceptionally fertile, 1980). THE

An outgrowth of the limestone Kentucky Mountains or the Eastern agricultural bounty is the Kentucky Coal Field. The latter development of the bourbon terms often include the Pine industry. Early Blue Grass Mountain Overthrust and its settlers had an excess of corn at associated basin and mountains as the end of the season and had no part of the region, but this convenient means of transporting discussion wi11 consider these the bulky commodity to Cincinnati areas as part of the Cumberland or other markets. A bulk reduction Mountains. The reason for this of the corn to distilled spirits separation is the independence of was a logical result. In 1789 speleogenetic processes at work in Reverend Elija Craig, of the development of caves and karst Georgetown in Bourbon County, in the two areas, and the great developed the recipe for a malt difference in the two areas' whiskey that now carries the name geomorphic history. of the county in which it was The Cumberland Plateau is a first distilled. Even today, State maturely dissected landscape where law requires that 51 percent of Missi ssippi an 1i mestones are the grain in bourbon must be corn, capped by more resistant thus imparting the unique taste to sandstones, shales, conglomerates, the beverage. Kentucky at one time and coal -bearing deposits of had over 400 distilleries Fennsyl vani an age. The maxi mum producing in excess of 88 percent development of karst features and of America's bourbon whiskey. caves is found along the Through corporate mergers and the Fottsville Escarpment. It is here necessity of large scale that the extensive caves of operation, the number has dropped Fulaski , Rockcastle, and Carter to 28 distilleries; all but 3 are counties have developed. in the Blue Grass. Experts in Especially important to the bourbon whiskey attribute the landscape's development have been excellence to the use of pure, streams like the Rockcastle River, natural limestone that , Buck Creek, emerge from the karst springs of Licking River, and Tygarts Creek. the area. Kentucky still produces These rivers have entrenched 75 percent of all United States' themselves through the bourbon (Raiti , 1980). Pennsylvanian sediments, exposing The karst springs were also the under1yi ng Mississippi an instrumental in the early 1imestones. settlement patterns of the Blue Few level areas are present on Grass. Cities such as Lexington, the plateau surf ace. The mature1y Harrodsburg, and Georgetown were dissected Rockcastle Conglomerate built at the sites of karst and Corbin are so springs in order to assure a dissected that only small ribbons steady supply of water for the of level interfluves are present. settlers. Even today, the city of Soils on these flat areas are Georgetown still gets its water re1at i vel y poor and support on1 y supply from the great flow of the barest, subsistence Royal Spring. agriculture. In the valley bottoms the agricultural production is Cumberland Plateau better, but the alluvial lowlands are severely constricted to small The gently southeastward meander bends on present and paleo dipping upland surface east of the drainage. The largest low1and Pottsville Escarpment and west of areas are related to shale the steep1y rising Pine Mountain outcrops that have been weathered is referred to as the Cumberland and eroded into bottom flats that Flateau. The Cumberland Plateau is support 1imi ted agriculture. What often referred to as the Eastern this region lacks in agricultural OVERVIEW OF THE GEOLOGY AND PHYSICAL GEOGRAPHY OF KENTUCKY 13

productivity is offset by coal activity. Oil and natural gas are also produced in this region. Through most of its extent the Pottsvi11 e Escarpment is a southeast-dipping cuesta capped by more resistant beds of Hockcastle Conglomerate of Pennsylvanian age. Near the Tennessee border the escarpment reaches 550 m, with a , combination of Hockcastle, Corbin, and Lee strata forming the cuesta front. To the north, near the Ohio River, the escarpment is known as the Waverly Escarpment, a low relief feature resulting from a pinching out of the Rockcastle conglomerate. Of geologic and geomorphic note in this area are and the sandstone arches. Cumber1 and Fa11 s developed where the Cumberland River crosses resistant sandstones and conglomerates and then falls to the level of in the lower Pennyroyal Province. The falls have retreated upriver from the Burnside area to where they are today. Not far from Cumberland Falls one can visit the beautiful sandstone arches and in the National Figure 6. Portion of Pomeroy- Forest. Further north, at Red ton Quadrangle showing the Red River State Park and at Carter River gorge and associated Caves State Park, one can also see natural bridge (from McGrain , good examples of natural bridges 1983, p. 35) . formed in the high1y eroded residual stream interf luves (Fig. Missi ssi ppi an 1i rnestones and 6). resulting in the formation of Hydrologic divides separate the several caves and arches. region into several cave and karst subregions. The most important in terms of the number of caves and the length of cave passages is the East of the Cumberland Plateau Pulaski/Rockcastle area where such is the zone of Pine Mountain or caves as Sloans Val ley, Coral, Cumberland Overthrust, whose Cave Creek, Hails, Big Sink, northern and western boundaries Precinct 11, and Goochland are terminate at the Russel Fork found. The caves are formed along Fault. This highland region the maturely eroded front of the contains the highest elevation in Pottsville Escarpment, where it is the State, 1,265 m on Big Black crossed by the Rockcastle and Mountain in Harlan County. The Cumberland drainage. The second western part of this province is major cave region is at Carter the southeastward-dipping strata Caves where Tygarts Creek has cut of the Lee Formation, which dorms through the caprock exposing the Pine Mountain. The eastern border 14 MISSISSIPPIAN PLATEAUS

is formed by the cuesta of the westward-dipping Lee Formation that forms Cumberland Mountain; with the Middlesboro Basin being a large synclinal valley between the two mountains. The whole area is a huge block 200 km long and 40 km wide that was elevated and displaced westward as much as 10 km from the southeast (NcGrain, 1983). The area has a diverse history of stream entrenchment and exhibits water gaps and wind gaps representing the effect of superposition and stream piracy. As in the Cumberland Plateau, the Mississippi an 1i mestones are over1ai n by Pennsylvanian shales, sandstones, conglomerates, and coal-bearing strata. Some of the most productive coal-mining in the State is on the elevated strata of the Middlesboro Basin. This basin is similar in composition and topographic expression to the 1ower Cumber 1and Plateau. Because of the overthrust, the region presented a formidable obstacle to transportation and settlement in pioneer days. Settlers were forced to traverse Figure 7. Topographic map of this area by crossing the famous Pine Mountain and nearby , a wind gap in Cumberland Plateau (from Cumberland Mountain. 4nother McGrain, 1983, p. 33) . topographic obstacle is the Breaks of the Sandy, where the Russel Fork of the Big Sandy River cuts area is best characterized by its through the northern end of Pine topographic expression. It is Mountain in a 300 m gash through divided into two major parts, the the Lee sandstones. This rugged upper Mammoth Cave Plateau ad the lower Pennyroyal Plateau. The landscape exemplifies the Pennyroyal Plateau is adjacent to topographic obstacles to the lower Blue Grass Region being development faced by the original separated from by The Knobs settlers (Fig. 7). it Cave and karst development is Region and Mu1 draughs Hi 11. The restricted to the thin exposures latter trends to the east and of limestone that outcrop on the merges with the Pottsville steeply dipping mountains. Because Escarpment and pinches out the of the great relief in the area, Pennyroyal on the eastern side of the Cincinnati Arch. The the deepest caves and the largest Pennyroyal extends on a westward pits in Kentucky are found in this arc, roughly following the district. Cumberland River through Kentucky and Tennessee. It then arcs Mississippian Plateaus northwestward to the junction of the Cumberland and Ohio rivers. On A1 though composed of Middl e to the west and north, the Pennyroyal Upper Mississippi an outcrops, this is bounded by the Dripping Springs OVERVIEW OF THE GEOLOGY AND PHYSICAL GEOGRAPHY OF KENTUCKY 15

Escarpment, which separates it from the Mammoth Cave Plateau. The Pennyroyal Plateau forms a corridor extending away from the Ohio River between the Dripping Springs Escarpment and the . The corridor continues as it swings back toward the Ohio River, skirting the Western Kentucky Coal Field from which it is separated by the Mammoth Cave Plateau. It was through this plateau corridor that many of the original settlers found an easy transportation route. The limestone derived soils are very fertile and have been extremely productive agriculturally, although they are less fertile than the Blue Grass phosphatic soi 1s. Much of the Pennyroyal Plateau exhibits extreme karstification, with numerous sinkholes dotting the landscape (Fig. 8). Sinkholes become less noticeable as one approaches the axial section of the Cincinnati Arch, where the karst-prone St. Louis and Ste. Genevieve formations have been stripped away revealing the less soluble Warsaw and similar formations. On the eastern side of Figure 8. Topographic map of the Cincinnati Arch, the karst the Park City Quadran~leof the landscape is once again obvious; a Pennyroyal plateau showing its 1arge number of sinkholes and high density of sinkholes (from karst features have developed on McGrain, 1983, p. 50) . the St. Louis formation. The Mammoth Cave Plateau is bounded on the outer margin by the Dripping Springs Escarpment. The between Elizabethtown and the Ohio escarpment is capped by sandstone River, but this is not being done overlying the St. Louis, Ste. in this volume. Genevieve, and Girki n f ormati ons Mammoth Cave is unsurpassed in that comprise much of the length because of the ideal cave-f orming limestone in the conditions for speleogenesis on Mammoth Cave area. At the inner the Mammoth Cave Plateau. The margin of this district is the permeable nature of the overlying higher Western Kentucky Coal Field sandstone permitted maximum water of Fennsylvanian rock. In this penetration of the limestone text, the Mammoth Cave Plateau is layers. Meanwhile, caprock divided into a Mammoth Cave Region continued to provide a stable and a Western Kentucky Karst ceiling for the area (Fig. 9). Region because of differences in Gently dipping toward the Green the drainage basins responsible River, which became entrenched for the karst processes. An f 01 1owing the Pleistocene argument could also be made for formation of the Ohio River, the the inclusion of another district plateau developed several 1evels 16 KNOBS

bottoms are often flat and broad (Fig. 10). Regional soils are of

.I,.. MAMMOTH CAVE k ,,L , ULhlhlll-

Western Kentucky Coal Field Figure 9. Mammoth Cave Region cross sections showing multi- The Western Kentucky Coal Field level development of the cave is a large structural basin and the resistant protective separated from the lower Mammoth sands tone caprock (from McGrain , Cave Plateau by the Pottsville 1983, p. Escarpment. The escarpment is 54). capped by the Pennsylvanian Caseyville Sandstone, forming a be1t of rugged hi11s rather than of cave passages. The thickness of the rock, the rock structure, and the factors just identified have resulted in the formation of a cave system that may some day exceed 900 km in length.

Knobs

The Knobs Region is a thin belt of conical knobs that surrounds the Blue Grass. The Knobs are erosional remnants of Mu1 draugh Escarpment to the west and the Pottsville Escarpment to the east. As the upland surfaces are eroded back, interfluve areas, protected by caprock of sandstone and shale, remain as isolated hills surrounded by a base level at the elevation of the Blue Grass. On the western side of the Cincinnati Arch the knobs are primarily Siluri an to Middle Devonian limestones, and on the east they are predominentl y shale. The region is, therefore, sandwiched between the Blue Grass Ordovician rocks and the younger Mississippian rocks of the Pennyroyal Plateau. The Knobs Region reaches its greatest width south of Louisville, where there is a gentle dip and no faulting. Figure 10. Topographic map of In the southern Blue Grass, the Junction City Quadrangle faulting is present and the dip is steeper, resulting in Knobs that showing the residual hummocky

are narrow in areal extent.- - terrain and the broad lowlands ~lthoughboth areas are rugged between the knobs (from McGrain, topographically, the stream 1983, p. 47). OVERVIEW OF THE GEOLOGY AND PHYSICAL GEOGRAPHY OF KENTUCKY 17

an imposing cliff face. This area Livesay, Onn, 1953, Geology of the is structurally a virtual Mammoth Cave National Park Orea: duplicate of the Eastern Kentucky Lexington, Kentucky Geological Coal Field of the Cumberland Survey, Series 9, Special Publi- Plateau, although the elevation is cation 2, 40 p. lower and the relief is not as great. The presence of sandstones McFarlan. 0. C.. 1943, Geology of and faulting has resulted in Kentucky: Lexington, University several hilly areas standing in of Kentucky. 531 p. relief against broad valleys f orrned from eroded shales. River McFarlan, fl. C., 1958, Hehind the valleys such as those of the Green scenery in Kentucky: Kentucky and Tradewater rivers were eroded Geological Survey. ser. 10, deeply prior to the Pleistocene, Special Publication 10, 144 p. and subsequently filled in with allu.vial deposits during ponding McGrain, Preston, 1954, Geology of caused by ice blockage of the Ohio the Carter and Cascade Caves River. The resulting broad flat area: Kentucky Geological Sur- areas are readily amenable to vey, ser. 9, Special Publication agriculture. The agricultural 5, 32 p. productivity of the area is in sharp contrast to coal mining McGrain, Preston, 1975, Scenic which results in ruinaton of geology of Pine Mountain in Ken- agricultural land. Few caves are tucky: Kentucky Geological Sur- found in this area because of the vey, ser. 10, Special Publica- lack of exposed limestone on the tion 24, 34 p. predominently Pennsylvanian rocks. McGrain, Preston, 1983, The geo- logic story of Kentucky: Ken- tucky Geological Survey, ser. 11, Special Publication 8, 74 p.

Moore, G. W. , and Sullivan, G. M., 1978, - the study of REFERENCES caves: Teaneck, New Jersey, Zephyrus Press. 150 p. Bogli, Alfred, 1980, Karst hy- drology and physical speleology: Raitz, K. B., 1980, The Kentucky Berlin, Springer-Verlag. 284 p. Blue Grass: Chapel Hill, North Carolina, University of North Jakucs, Laszlo, 1977, Morphogenet- Carolina, Studies in Geography, ics of karst regions: New York, Halsted Press, 284 p. No. 14, 151 p.

Jennings, J. N., 1971, Karst: Lon- Sweeting, M. M.. 1972. Karst land- don, MIT Press. 252 p. forms: Londdn, MacMillan, 362 p. Chapter 2 CAVES OF KENTUCKY Angelo I. George A. I. George Consultants 1869 Trevilian Way Louisville, Kentucky 40205

The caves of Kentucky have been exploit their economic worth in important in the history of set- the early 1800s. This inventory tlement patterns for both the was followed by academic inquiries American Indians and the pioneers. into the distribution of caves Many of the cave entrances served because of their unique fauna, as living quarters to the Indians. flora, and archeological signif i- The Indians explored some of these cance. The contemporary inventory caves and exploited the of caves in Kentucky began with resources such as water, , early efforts by the National epsomite, flint, and . The Speleological Society, followed by best examples are Mammoth Cave, local and individuals Salts Cave (Edmondson County) and collecting and mapping caves in Great Wonderland Cavern (Hardin this state. County). By the middle part of the 18th century, pioneers were moving In 1806 a Lexington saltpeter- westward into a land called Ken- gunpowder entrepreneur by the name tucky. Like the Indians. the pio- of Samuel Brown wrote of 6 caves, neers used caves for shelter (Cave giving their name, location, some Hut, Muhlenburg County); they too notes on each cave, and a detailed exploited the mineral resources description of one of them. He found in the caves, especially knew of others, but concentrated water, epsomite, copperas, and on the ones that he owned in saltpeter. Caves have been used as present day Rockcastle, Jackson, storehouses for foodstuffs and Hardin counties. (Breeding Saltpeter Cave, adair County), mushroom farms In 1817, Luke Munsell was hired (Constantine Saltpeter Cave, by the state government as chief Hardin County), and livestock surveyor and directed to produce a quarters (Hidden River Cave, Hart high quality, detailed map of County). During these early times, Kentucky. Ry using existing maps few records of caves are known. and his prior survey field work, What is known about this early coupled with new field period is based upon contemporary investigations, Munsell by 1818 investigations into the history had published his "Large Map of Kentucky." This was the first and archeology of the site. really accurate map of the State The first inventory of caves in which showed the positions nf Kentucky sprang out of a need to cultural and natural features in CAVES OF KENTUCKY 19 their true spatial relationships century historians followed their with one another. Munsell took it outline of cave deacriptons. upon himself to include remarkable Throughout the 19th century, the geographic places. The completed fame of Mammoth Cave dominated the map thus shows a number of karst cave scene with a host of features, sinking streams. descriptions and maps published on springs, dry stream channels, and this famous wonder of the world. cave entrances. No other single There would be sporadic references map in the State was to include 50 to nearby caves, such as Dixon, many karst features up to the time Long. Short, Hundred Domes, James, of our modern topographic mapping Diamond, Hidden River, and Mammoth program. The first real catalog of Onyx caves. Occasional1y, a caves and their distribution was travelogue would describe other published by Constantine S. caves in the State. Alexander Rafinesque in 1832. This catalog Wilson (1810) wrote of is based upon his tenure at Rleedinghearts Cave (Warren Transylvania College (Lexington. County), and the anonymous Kentucky) between 1818 and 1826. (attributed to Constantine S. He was the first to examine the Rafinesque, 1820) account of caves' geological and zoological Russell Cave (Fayette County) made content. He devised seven for a fine addition to the speleo different classif icatons of caves: history of the State. But, all in cliff caves (rock or rock all, caves outside the Mammoth houses), fissure caves, sinking Cave Region seem to have little to caves, spring caves, crater or offer in comparison. funnel caves, saltpeter caves, and Between the time of Rafinesque stalactical caves. He gave names, and Alpheus S. Packard there is a locations, and physical hiatus with little regional descriptions of eight caves in collection of cave sites. Packard Kentucky. Four of the caves conducted an extensive inventory studied occur in the Inner Blue of cave life from the caves of Grass, two caves in the Cumberland Kentucky and published his Plateau of Rockcastle County, and findings in 1888. He mentioned two caves in the Wississippian more than 50 caves from site data Plateau of Edmonson County. and collections by Nathaniel S. Kentucky historians have played Shaler and his Kentucky Geological a key role in the preservation of Survey in 1875. Arthur M. Miller accounts of early cave (19191, State Geologist, listed 16 explorations and discoveries. representative caves and 5 areas Lewis Collins (1847), followed by where caves occur. Gerard Fowke his son, Richard Collins (1874), (1922) describeed 39 caves in the in their "History of Kentucky," State as part of his midwest recorded the location and physical archaeological investigations. descriptions of caves by county. From 1923 to 1957, William D. Data were gathered without credit Funkhouser and William S. Webb from old newspaper accounts. explored, excavated, and cataloged travelogues, and a smattering of caves as part of a Statewide scientific publications. Cave inventory of archeological sites descriptions selected were usually in Kentucky. of the kind with bottomless pits The massive publicity resulting (Frenchman Knob Pit, Hart County) from Floyd Collin's entrapment in or lost Roman treasure caves Sand Cave, Edmonson County. (Love11 Cave, Muhlenburg County). generated a demand for more If the cave contained animal newspaper articles on caves, and bones, especially mammoth bones, not just caves in the Mammoth Cave the finding of such in a cave was vicinity, but caves all over a favorite vignette of the Kentucky. There was a plethora of Collins'. All succeeding 19th newspaper articles published 2 0 GEOGRAPHIC DISTRIBUTION OF CAVES IN KENTUCKY between 1925 and 1935. The maximum Province in the southeastern amount was in 1925 during the United States. The State can be Collins incident and just after divided into six major geomorph.ic his death. regions that closely align In 1943, the fledgling National themselves with prominent Speleol ogical Society undertook lithological, structural. the cataloging of all known caves hydrological, and geomorphological in the United states and the changes in the topography (Fig. world; Robert Morgan began this 2). The geomorphic regions are task and produced the first modern (from east to west): Eastern catalog of caves. By 1944, he had Kentucky Coal Field, Knobs, Blue inventoried 102 caves in the State Grass, Mississippian Plateaus, and the list was documented with Western Kentucky Coal Field, and an extensive bibliography. From Jackson Purchase. Caves are that time, grottos, surveys, and developed in Paleozoic carbonates individuals have maintained ranging in age from Middle extensive catalogs of caves in Ordovician through Late Kentucky. Mississippian. We1 1 developed Collection efforts over the pseudocaves occur in Lower last 20 years has yielded 3,770 Pennsylvanian conglomerates and caves and 1,057 cave maps spread clastics. No caves hae been out in 87 of the 120 counties in reported in anything younger than Kentucky (Fig. 1, Table 1). The Early Pennsylvanian in age. gridwork display in Figure 1 was Structural development of the constructed by plotting cave Cincinnati Arch coupled with entrances on 2.5 minute centers of sediment downloading of the latitude and longitude. Eastern and Western Kentucky Coal basins at the close of the GEOGRAPHIC DISTRIBUTION Paleozoic Era set the stage for OF CAVES IN KENTUCKY the present distribution of caves in Kentucky. Major landscapes as Kentucky is contained in parts we see them today have been in of the Coastal Plain, Interior Low existence since the end of the Plateaus, and Cretaceous Period. Generally, cave

Figure 1. Distribution of caves in Kentucky, plotted on 2.5-minute centers of latitude and longitude. CAVES OF KENTUCKY

Table 1 Listing of Caves in Kentucky by County

Total No. Caves 3770 Co. with Caves 87 No. Mapped Caves 1057

.------NO. NO. HAP NO. NO. MAP COUNTY CAVES CAVES COUNTY CAVES CAVES

Adair 35 6 Laurel 1

Anderson 18 Lee 43 Allen 20 7 Letcher 27 Barren 135 43 Livingston 14it Rreckinridge 425 7 1 Logan 55* Be1 1 B -.-. Lyon 9+ Roone 8 Madison 27 Bourbon 12 6 Magof f in 2 Boyde 1 Marion 9 Roy 1 e 8 Mason 1 Rullett 30 30+++ McCreary 14 B~ttler 10 Meade 212 Cal dwell 39rc 25 Menif ee 4 Carter 187 36 Mercer 37 Carroll 1 Metcalf e 19 Casey 6 Monroe 22 Christian 24 15 Morgan 1 Clark 9 1 Montgomery 4 Clinton B 3- Muhlenburg 3 Crittenden 25 B Nelson 2 1 Cumberland 16 4 Nicholas 1 Edmonson 173 44 Ohio 4 Elliott 2 1 Oldham 13 Estill 37 Owen 5 Fayette 72 2 1 Pike 1 Floyd 2 Powel l 37 Frank1in 28 1 Pulaski 148 Garrard 10 Hockcastle 181 Grayson 4 1 9 Russel 1 10 Green 71 B Scott 12 Greenup 1 She1by 3 Handcock 1 Simpson 48 Hardin 297 62 Tayl or B Harlan -7 Todd 18 Harr ison 1 Trigg 66 Hart 207 74 Trimble 10 Henderson 2 Union 2 Henry 23 3 Warren 198** Hopkins 2 Washington 1 Jackson 134 10 Wayne 94 Jefferson 109 32+xit Whitley 1 Jessamine 50 12 Wolf 7 tinott 1 Woodf ord 40 Larue 23 Miscel 1aneous 3 *John Mylroie, personal communication, 1984. **Central Kentucky , Bull. No. I. ***Phillip Di81asi , personel communication. EASTERN KENTUCKY COAL FIELD REGION

Figure 2. Six major geomorphic regions of Kentucky. development sought out Pennsylvanian cyclothemic structurally favorable areas where sandstones, shales, coals, there was a maximum amount of siltstones, and limestones on the relief between the point of downdip side (Fig. 2). The basin recharge and the point of can be subdivided into three discharge along base level sections: Cumberland Mountain, streams. The great caves of long Kanawha, and Cumberland Plateau. extent are all found adjacent to Caves are found in all three structurally induced Chesterian sections. The majority of the (Mammoth Cave System, Edmonson caves are developed within the County) and Pottsville escarpments Cumberland Mountain and Cumberland (Cave Creek System, Pulaski Plateau sections. County) or along the crests of anticlines (Hayes Cave, Grayson Curnberland Mountain Section County) and synclines (Big Bat Along the southeastern boundary Cave. Breckinridge County). of the Eastern Kentucky Coal Field Frequency of cave entrance is the Pine Mountain Overthrust occurence, passage size, and Fault. This fault has formed a groundwater yield decreases in the synclinal trough flanked by two direction of recharge boundaries. parallel mountains, Cumberland and This generalization can be seen in Pine. The northwestern slope of the north-central Kentucky karst Fine Mountain contains the and is consistent with majority of caves in this Swinnerton's (19.72) theoretical locality. Here, Mississippian speleogenesis model. carbonates have been overthrust to the northwest out and on top of Eastern Kentucky Coal the Kanawha section of Field Region Fennsyl vanian clastics. The The Eastern Kentucky Coal Field carbonates are extensively Region is a synclinal structural fractured and caves have formed basin filled with Nississippian along the strike and Pine carbonates and clastics on the Mountain. Pine Mountain stretches updip side, and overlain with for more than 160 km through KENTUCKY

Kentucky and contains over 91 m of within the northwestern portion of carbonates. This karst area the Kanawha section. Nearly 200 represents probably one of the caves are known from this area. greatest untapped caving Many of these caves are long localities in the State. There is braided networks or caves. much potential for big cave. deep Rest known are Bat Cave, Saltpeter pits, and systems to keep cavers Cave, Burchetts Cave, Carter City busy for years. The most famous Connection, Iron Hill System, and caves are Linefork Cave, Icebox the Cascade System. The highest Cave, the Payne Gap Cavern, Payne density of caves occur along the Gap Water Cave, and Angel Cave. axis of a syncline, truncated by The Colehole is a 73-m-deep pit deeply entrenched Tygarts Creek. with no passage at the bottom. This structural setting provides Sand Cave is a pseudokarst feature for maximum rock fracturing along formed in the Lee Sandstone. the axis of the syncline, giving Carbonate cave passage rise to network cave development. configuration is similar to that Cave entrance distribution in seen in the Valley and Ridge areas Carter County shows a of . northeast-southwest stike orientation along the Pottsville Kanawha Section Escarpment. Major streams follow The Kanawha Section is the this trend toward the Ohio River. heartland of the Eastern Kentucky Pseudokarst in the form of Coal Field. It is a paradise for natural bridges, rock shelters, the coal and oil barons, and a lighthouses, and dolines are well bane for the caver, with developed adjacent to valleys impossible roads, rugged terrane, breaching the Pottsville and few caves. Few caves are Escarpment. These features have memorable, and the geologic formed in the Lee Formation, a section is full of Pennsylvanian sandstone-conglomerate of Early clastics, coals, shales, and thin Pennsylvanian age. Best recognized carbonate units. There are less are the natural bridges and than 20 caves indexed from this pseudokarst caves of the Red River section in Kentucky. Gorge in Powell County and the in McCreary County. Cumberland Plateau Section Inspection of cave entrance Stretching for more than 190 km density reveals a dominant across the east central part of northeast-southwest trend parallel Kentucky is one of the paramount to the strike of the Pottsville cave and karst localities in the Escarpment. Dominant cave passage State. The locality is bounded on trends are to the northeast and the west by the northwest. This same pattern is Escarpment and on the east by the repeated in most of the major Pottsville Escarpment. More than caves throughout the State. 840 caves are known from this Structural upwarping of the area. Caverns of every possible Cincinnati Arch and development of description can be found in this the Appalachian Mountain System section, whether large, dry are responsible for fostering the walking passage. or wet and muddy original fracture pattern, visible slush tube. Representative caves today in this dominance of cave are: Sloans Valley System, Cave passage trends. Creek System, Goochland-Smokehole Complex, Great Saltpetre Cave, Mississippian Plateaus Region Triple S Cave, and Eureka Cave. The Mississippian Plateaus Ry rights, the Carter County Region (Fig. 3) is the largest karst should be in the Kanawha single continuous sequence of section; rather, it exists as a exposed fractured karsted finger of Hississippian carbonates carbonate rocks in Kentucky. The MISSISSIPPIAN PLATEAUS REGION

1-! """4~I PL'AIN

Ease from GeoIog1~Map of Kentucky Series IX 1954 Revised from Geologic T E N N E S S E E Ma 9 of hcntu$hy doted 1927ond 19$9 by W.R. Jillson 0 ?s 50 MILES

Figure 3. The Mississippian Plateaus Region, which contains 2,125 known caves, including Mammoth Cave.

region contains the largest Just north of Mammoth Cave is a compliment of caves with 2,125 Pennsylvanian conglomerate and known: the largest cave system in sandstone f i 1led fossi 1 river the world, Mammoth Cave; and a valley. Contained within the collection of lesser systems that Prownsville Channel are are among the great long caves in pseudokarst caves, dolines, the world: Whigpistle, Hicks, Pig springs, and . Pat, Lost River, Lisanby, Big Representative are: Big Spring, Sulphur. and Gradys. Holley, and Lines caves. Deep There is a distinct regional shafts have formed as a result of clustering of cave entrances along initial pseudokarst development the Chester Escarpment from followed by interstratal Prandenburg in Meade County karstification at through Bowling Green in Warren clastic-carbonate contacts. Good County. Cave entrance density examples can be seen at Frenchman declines in both directions away Knob Pit and Lost Hound Pit, Hart from the escarpment. The largest County. of the cave systemsare all found In the far northwestern part of beneath the Chester upland or the region is the Western Kentucky below outliers from this landform. Fluorspar District. The Rough Large graded trunks and bedding Creek Fault Zone has intensively plane tubes are the typical cave fractured this part of the State. passages. These passages occur in It is the site of hydrothermal the Middle Mississippian carbonate activity. Karstification has units. In contrast, the character occurred along the faults, joints, of the cave passages in Chesterian and bedding planes. These features rocks is determined maze. have been filled in with CAVES OF KENTUCKY

from the hydrothermal activity. sub-regions: Inner Blue Grass, Over 50 caves are known from this Eden Shale, and Outer Elue Grass local ity. The re1ati onshi p between (Fig. 4). The geomorphic hydrothermal activity, and cave subdivisions are defined upon development is a subject for lithologic and structural controls further investigation. of the Cincinnati Arch and the upwarped Lexingtan Dome. The Western Kentucky Coal overall picture is one of Field Region concentric bands of contrasting Like the Kanawha Section in the rock types. The Inner Rlue Grass Eastern Kentucky Coal Field contains Middle Ordovician Region, cave development is sparse carbonates and shales; the Eden in this region of Kentucky. The Shale Eel t contains Upper few caves that are found are Ordovi cian shal es; the Outer Hlue mostly clustered along the strike Grass contins Upper Ordovician and of the Pottsville Escarpment, and Silurian-Devonian carbonates and rapidly decrease in number shales. westward into the interior of the structural basin. More than 60 Inner Blue Grass Subregion caves are known from this region. The cave and karst area of the Tectonic activity along the Inner Rlue Grass is a triangular Rough Creek Fault Zone has area centered in Fayette County. fostered the conditions for cave Its base is just east of Lexington development in central Grayson and its apes extends into Woodford County. More than two thirds of County. The highest density of the caves known in the county raves occurs in central Woodford, occur on or adjacent to faults. A Fayette and Jessamine counties. In large scale , called The southern Fayette and in Jessamine Sink, is developed along the Rough counties, cave and pit entrances Creek Fault Zone in Missi ssippi an show good para1lel a1 ignment along carbonates. This polje is more the strike of master streams. than 40 km west of the nearest Few caves are developed east of surface outcrop of Middle the Lexington Fault System, which Mississippian rocks. is the eastern limits of the

Jackson Purchase Region The Jackson Purchase Region in western Kentucky is part of the northern extension of the Mississippian Embayment. The embayment was filled in with continental sediments at the close of the Cretaceous Period. These sediments have buried the western end of the Mississippian Plateaus Region. There are no enterable caves in this section of Kentucky: but, well drilling activity has recorded karren topography with 15 to 30 m of relief, buried beneath more than 305 m of sediment. Karst cavities or better yet, caves, occur below the karren zone of Faducah. Figure 4. Divisions of the Blue Blue Grass Region Grass Region, showing the con- The Blue Grass Region can be centric bands of contrasting divided into three main rock types. EDEN SHALE BELT SUBREGION

Lexington karst. There is a approached, basal carbonates of recurrence of caves in central Ear 1y Mississippian age are Bourbon and Clark counties. The present on top of the knobs, and caves and karst of Clark County this is where the caves are are in one of the few areas where 1ocated. the Eden Shale is thin, and caves have formed in the underlying CONCLUSION carbonates adjacent to stream valleys. This situation also holds This discussion has only been a true for the karst in the vicinity brief overview of the distribution of Harrodsburg, Mercer County. of caves in Kentucky. There are Representative caves in this many caves known in addition to locality are: Hryants, Russell, the ones used in the compilation Phelps, and Foggs. Over 260 caves of this report. Continuous effort have been inventoried in this during the last 20 years has locality. allowed the inventory of 3,770 caves in Kentucky. If the past is Eden Shale Belt Subregion any indication of the future, the Generally, the Eden Shale Belt next 20 years looks very bright. is devoid of cave development, except where the shale aquitard is ACKNOWLEDGMENT thin and the underlying carbonates are in close proximity to base Much thanks to all individuals level drainage. who contributed cave data over the years. Outer Blue Grass Subregion The Outer Blue Grass marks the REFERENCES recurrence of high density cave and karst development. As in the Anonymous, 1820, Russels Cave: Inner Blue Grass, cave entrances Western Review, v. 4, p. 161- tend to cluster near base level 163. drainage and rapid1y decrease toward drainage divides. There are Brown, S., 1809, A description of two areas where most of the caves a cave on Crooked Creek, with occur. Several caves, such as remarks and observations on Adams Saltpeter Cave in Madison nitre and gunpowder: American County are found in the southern Fhi losophical Society Transac- part of the subregion. The tions, v. 6, p. 235-247. majority of the caves are related to the deep entrenchment of the Coll ins, L. , 1847, History of Ken- Ohio River and its tributaries tucky: Maysville, Kentucky, 560 a1ong the f 01 1owing counties: P - Henry, Oldham, Trimble, Shelby, Jefferson, Bullitt and Nelson. Collins, R., 1874, History of Ken- About 280 caves are known and most tucky: Covington, 2 vols. are small in lateral extent, Fowke, G., 1922, Archaeological rarely exceeding 150 m in length. Investigations: Smithsonian In- Caves tend to have big entrances stitute Bureau of American Eth- whose passages degrade into nology, Bulletin 76, 204 p. crawl ways. McFarlan, A. C., 1943, G~ologyof Knobs Region Kentucky: Lexington, University The Knobs Region contains about of Kentucky, 531 p. 15 caves. Most of the knobs are made up of siltstone and shale, Miller, A. M., 1919, The Geology which are not conducive to cave of Kentucky: Kentucky Department development. As Muldraughs of Geology and Forestry, ser. 5, (Highland Rim) Escarpment is Bulletin 2, 392 p. CAVES OF KENTUCKY 27

Morgan, R. E., 1943, Caves in Packard, A. S., 1888, The cave world history: National Speleo- fauna of North America: National logical Society Bulletin, no. 5, Academy of Science, v. 4, pt. 1, p. 1-16. 156 p.

Rafinesque, C. S., 1832, The caves of Kentucky: Atlantic Journal Morgan, R. E., 1944, Additions to and Friends of Knowledge, v. 1, "Index of all known caves of no. 1, 27-30. the worl dl': National Spel eol og- p. ical Society Bulletin, no. 6, Swinnerton, A. C., 1932, Origin of 29-33. p. 1 i mestone caves: Geological So- ciety of America Bulletin, v. 43, p. 663-693. Munsell, L., 1818, & map of the State of Kentucky from actual Wilson, A., 1810, Letter from surveys, etc. : Phi 1adel phi a, Nashvi 11e: The Port Folio, v. State of Kentucky. 4, p. 310-321. Chapter 3 THE INNER BLUE GRASS KARST REGION John Thrailkill Department of Geology University of Kentucky Lexington, Kentucky 40506

This discussion of the Inner Scott counties area), and D. R. Blue Grass Region of central Ken- Gouz ie (Walnut Hi11 area) . None tucky has been extracted from of what follows would have been chapters A and F of Thrail kill and possible without their efforts. A others ( 1982) . On1 y minor changes description of each of these have been made to correct typogra- areas, as we1 1 as detailed infor- phic errors and delete references mation on the dye traces made and to other chapters and omitted por- spring discharges determined, is tions of chapters A and F. found in Thrailkill and others The principal method used to (1982). study the area was water tracing, Portions of the work described and a description of the dyes and in the present report have already techniques used, as well as rela- appeared (McCann, 1978; Thrailkill ted laboratory studies (Society of and Troester , 1978; Thrai 1ki 11, Dyers and Colouriets, 1971; 1980: Thrai 1ki 11 and others, Buinlan 1977; Buinlan and Howe, 1980; Spang1 er and Thrai 1ki 1 1 , 1977: McCann, 1978; Byrd, 1981; 1981; Thrailkill and others, 1981: Buinlan and Ewers, 1981; Spangler, Spangler, 1982) or have been 1982; Spangler and others, submit- accepted for publication ted for publication) are described Thrailkill and others, accepted in Thrailkill and others (1982). for publication). The focus of work in the region Major support for this work was has been on the geol ogi cal , water provided by the Office of Water supply, and environmental aspects Research and Technology (now with- of the karst and its , and in the United States Geological a1though valuable information on Survey), United States Department these subjects has been obtained o+ Interior, as authorized by the from studies in caves (as will be Water Research and Development Act discussed), such studies have been of 1978, F'ublic Law 95-467. Addi- incidental to the regional study tional funding was from the to date. No listing or cave loca- McFarlan Fund of the Department of tions will be presented, both Geology, University of Kentucky; because such a listing would be Dames and Moore: the Georgetown incomplete and because of the Municipal Water and Sewer Service; sensitive 1andowner re1ati onships the National Speleologi cal that exist in many area5 in the Society; and the Institute for region. Mining and Minerals Research, and The information that has served the Research Committee of the as the basis of the discussion University of Kentucky. The that follows has large1y been support of these agencies, as well derived from area studies in por- as the landowners of the region tions of the region outlined on and the many others who have Figure 1. The field investi gat ion5 assisted the research is grate- in these areas were conducted by f ul1 y acknowl edged. M. R. McCann (Northeast Woodford Work has, and is, continuing in County area), M. W. Hooper, Jr. the 3 years since the discussion (Mercer County area), L. E. that follows was written. Some of Spangler and J. W. Troester the research is described in Eyrd (northern Fayette and southern and Thrai 1ki 11 (198.:) , Hooper and THE INNER BLUE 'GRASS KARST REGION 29

......

/ ......

......

......

...... '....,......

...... , . . , ..,. . # ..,. . , ... ..,... Danvills 2 ...... :::::::::::::::::: .., ...... INNER BLUEGRASS ;:; . . , ...... KARST REGION ... BOY LE ...... L...... 'LY ...... Figure 1. Map of Inner Blue Grass Region. Coverage of larger scale maps in Thrailkill and others (1982) is also shown. THE INNER BLUE GRASS REGION

Thrai 1ki 11 ( 1983) , Thrai 1ki 11 generally considered to be (1983) , Thrai 1ki 11 (1984) , residual. The entire region is Thrailkill and Gouiie (1984), and south of the area modified by Thrai 1ki 11 (accepted for Pleistocene glaciation. The pub1ication) . present populaton is in excess of 350,000, of which more than half INNER BLUE GRASS REGION is concentrated at Lexington, the second largest city in Kentucky, The Inner Blue ~rassRegion is which lies near the center of the an area of 5,600 km2 in region (Fig. 1). central Kentucky. It is largely a gently rolling upland at an Geologic Structure altitude of 250 m, with generally less than 50 m of relief, that has The region occupies the area been termed the Lexington where carbonate rocks of Middle Peneplain (Jillson, 1961). Most of Ordovician age have been exposed the streams which drain the area by erosion on the crest of the are on the upland, but the Cincinnati Arch, a regional Kentucky River, that crosses the structural feature of the eastern region, has incised a gorge more United States. Regional dip is than 100 m deep. Altitudes range generally away from the highest from 350 m in the southeastern point on the arch in Jessamine portion of the upland to 130 m County (Fig. 1) in all directions along the Kentucky River where it except to the southeast, where the leaves the region in the rocks have been down faulted. northwest. Regional dip is gentle (on the Although the streams on the order of 10 m/km), and the beds upland surface appear to provide seen in outcrops generally appear normal surface drainage, numerous near1y horizontal. karst landforms (especially The southeastern boundary of sinkholes) are present, and the region follows the Lexington portions of the region, some with Fault System in the south and the areas in excess of 10 km2, intersecting Kentucky River Fault have no surface drainage. The System to the east (Black and outlines of the region (Fig. 1) others, 1977). The eastern and were defined by including within southern sides of these fault its boundaries all 7.5-minute systems are downdropped, and quadrangles (1:24,000) that depict unkarsti f ied Upper Ordovician at least one by limestones and shales cover the topographic contours (interval 3.0 Middle Ordovician carbonates. to 6.1 m) in rocks of Middle There are a few areas of Ordovician age. The Inner Blue substantial faulting within the Grass Region is both region, such as the Switzer Graben geographically and in Scott County and the extension stratigraphically distinct from of the Lexington Fault System to another extensive karst area (a the north. There are also a number portion of which has been termed of short, high-angl e faults and the Central Kentucky Karst) in mineralized veins. Mississippian rocks, as we1 1 as from smaller karst areas in Stratigraphy Kentucky in Upper Ordovician and The boundary of the region Silurian rocks. approximately coincides with the The mean annual precipitation depositional or fault contact of is about 1,150 mm, fairly evenly relative1y pure Middle Ordovician distributed throughout the year. carbonates with the overlying, Mean July and January temperatures thinly interbedded Upper are about 25 and (I0, Ordovician limestones and shales. respectively. The regolith is The overlying limestone and shale often a meter or more thick and is sequence has been designated the THE INNER BLUE GRASS KARST REGION

Clays Ferry Formation, and the In a study of the Lexington underlying carbonates are, from Limestone in Franklin County, highest to lowest, the Lexington Fisher (1968) found that the Limestone, Tyrone Limestone, maximum insoluble content of the Oregon Formation, and Camp Nelson Gri er and Tang 1ewood 1i mestone Limestone. members was 15 percent and averaged less than 5 percent. His A11 of the area studied to date data also indicate that the has been in the lower portion of maximum content of insoluble the Clays Ferry Formation and the minerals in units general 1y upper two-thirds of the Lexington considered argillaceous (Macedonia Limestone. The lower third of the Eed and Brannon Member) was on1 y (including the 25 perrent, and that lithologies Logana and Curdsville members) and usually described as shales are the three formations below Ft are usually more than 50 percent exposed only in the gorge of the calcite and . Kentucky River and the lower reaches of its tributaries, and Cressman 11973) calculated underlie areas not yet normative mineral percentages investigated. Furthermore, it is based on chemical analysis of believed the subsurface 15-cm core segments from the Clays circulation of meteoric water Ferry Formation and the within the area studied does not Millersburg, Erannon, Tanglewood extend into these units; therefore Limestone, and Grier Limestone these units will not be further members of the Lexington considered. Limestone. Analyses were performed on five core segments selected The principal lithologic random1y from the core avai lable characteristic of hydrogeologic for each of the five units. The interest in the Lexington mean quartz plus clay content Limestone and overlying Clays calculated for the Grier and Ferry Formation is the amount of Tanglewood limestone members was 8 insoluble material in the latter percent and 5 percent and in units of the former. This respective1y. For the remaining factor has been considered a major three units (considered control in the development of argillaceous) these amounts were: solution openings by most earlier Erannon Member, 38 percent; workers (Hamilton, 1948, 1950; Millersburg Member, 35 percent: Palmquist and Hall, 1961; Mull, and Clays Ferry Formation, 44 1968; Faust, 1977). Stratigraphic percent. descriptions of the Clays Ferry Although dolomite is present in Formation and the various subunits most of the units, especially the of the Lexington Limestone more argillacious ones, it accompany the pub1ished geological generally occurs as isolated quadrangles of the area studied ( rhombs. Fisher (1968) found the Black, 1964, 1967; Cressman, dolomi te-calci te ratio to be 1964, 1967, 1972: Kanizay and generally less than 0.2 and to Cressman, 1967; Miller, 1967; exceed unity only in one thin MacRuown and Dobrovolney, 1968; (less than 1 m) bed in the Grier Pomeroy, 1968, 1970; Cressman and Limestone Member. The normative Harbar, 1970; Allingham, 1972). mineralogy calculated by Cressman These descriptions are believed to (1973) yields mean values of this be based generally on hand ratio to be 0.1 and 0.17 for the specimen examination and usual 1y Grier and Tanglewood limestone state the approximate percentage members, respectively: the three of clay, chert, and other argillacious units he examined insoluble components, as we1 1 as range from 0.23 to 0.46 (above). note the occurrence of minerals The stratigraphic nomenclature such as dolomite and apatite. used on the various geologic maps 32 PREVIOUS HYDROGEOLOGIC INVESTIGATIONS

is not always consistent, and the in Table 1. Subdivisions of the terminology of Cressman ( 1973) Perryville in the Mercer County will be used in this report. area (i.e., Cornishville and Except for the Clays Ferry Salvisa beds) and the thin pure Formation and Nillersburg Member, Devils Hollow Member within the all of the argillaceous units (11 Tanglewood in the northeastern units of the Lexington Limestone Woodford County area have been except the Clays Ferry) are less omitted. (actual1y considerably less) than 6 m thick. The delineation of the Previous Hydrogeologic Investigations various units is based on lithology, and the units show A number of hydrogeologic complex gradational and studies of the Inner Blue Grass intertonguing relationships, which Region have been publ ished. The often result in multiple earliest of these, by tlatson occurrenres of a unit in the (19091, dealt with the larger Blue stratigraphic section. Grass Region, which includes In the northeastern Woodford extensive non-karst areas outside County, northern Fayette and the Inner Blue Grass Region. He southern Scott counties, and presented data on a number of Walnut Hill areas, the relatively in the present study area pure Tanglewood and Grier (e.g., 48 in Fayette County, 30 in limestone members make up most of Scott County, and 20 in Mercer the section. The argillaceous County , but with such general Millersburg Member, Greendale locations that they could not be Lentil, and Stamping Ground Member utlized in this study. His occur within the Tanglewood, the discussions of the Frannon, and Cane Hun members at are general and lack conclusions or near the Tanglewood - Grier regrading controls of groundwater contact, and the Macedonia Fed occurrence and movement in the occurs within the under1ying Inner Flue Grass Region. 4lthough Grier. In the Mercer County area, he mentioned a trace to a spring two relatively pure units overl'ie with oi1 and NaCl , and that NaCl the Tanglewood, and only two of was used in an examination oS the argillaceous units within the Royal Spring (Matson, 1909, p. Lexington Limestone are present. 80-81). he gives no location These re1ati onshi ps, which are inf ormati on. The on1y publ ished considerably simplified, are shown information on water tracing in

Table 1.--Stratigraphic Units in the Study Area. All (Except Clays Ferry Formation) are Units of the Lexington Limestone. * Indicates Unit Present Only in Mercer County Area: **Indicates Unit not Present in Mercer County Area.

LIMESTONE UNITS flRGILLACEOUS LIMESTONE UNITS

Clays Ferry Formation Mi 11ersburg Member*+ Sulpher We1 1 Member* Greendal e Lentil** Perryvi1 le Limestone Member* Stamping Ground Member Tanglewood Limestone Member Erannon Member Grier Limestone Member Cane Run Member** Macedonia Bed THE INNER BLUE GRASS KARST REGION I the region prior to the present Hal 1, 1960~1. Vari ati ons between study was presented by Ji11 son the assessments are probably due ! 19451, who established flow to differing evaluation criteria connections in the Roaring Spring and the density of well control. ground water basin in Woodford Palmqui st and Hal 1 's map (1960~), County. o+ the same four counties studied Hami 1ton ( 1950) reported an by Hamilton (1950) is apprently inventory of 964 wells in a four based on 64 wells and 31 springs, county area (Rourbon, Fayette, as opposed to the 964 wells listed Jessamine, and Scott). Although he by Hami 1ton. Thei r summary states lists the total depth of all but a that 35 wells and springs were few of these, he reports water inventoried in each county and levels in only 56 and hence could that water levels were measured in not prepare a map of the most we1 1s (Palmquist and Hal 1, potentiometric surface. He states 1961, p. 3, 151, but they give that only about one out of five neither the water level nor a map wells drilled is productive of the potentiometric surface. (Hamilton, 1950, p. 47-48) and The summary (Palmqui st and concluded (also in Hamilton, 1948) Hall, 1961) covers the entire Blue that solution porosity is limited Grass Region, and it is difficult to a depth of 25 meters, that such to separate their conclusions on porosity is developed mai nly a1 ong the Inner Elue Grass Region from joints and is greatest in the largely unkarstif ied areas topographically low areas. He that surround it. They appear to states that argillaceous limestone ascribe differences in well yields units within the Lexington in the Inner Blue Grass Region Limestone play a major role in more to topographic posi tion than that they severely inhibit the to stratigraphy (reflected in downward circulation of meteroic their hydrogeologic maps) , which water and hence retard the is more or less the reverse of development of solut i on porosit y Hamilton's (1950) criteria. They in the rocks that underlie them. also state that less than half of His maps, which delineate areas of the wells drilled in the bedrock high, intermediate, and low are successful (Palmqui st and probability of obtaining a Hall, 1961, p. 21). satisfactory yield and quality of Henderson and Kri eqer ( 1964) groundwater, are apparently based presented a summary of the mainly on stratigraphy. geochemistry of waters of the A series of hydrogeologic maps entire Elue Grass Region. A brief covering the Inner Blue Grass report and map on the hydrogeology Region (Hal 1 and Palmqui st, of Fayette County by Hopkins 1960a-d; Palmqui st and Hal 1, ( 1966a) expl dins groundwater flow 1960a-c) were issued as part of a in terms of regional and local statewide project, and a potential gradients controlled discussion of the hydrogeology of mainly by topographic factors, and the larger Blue Grass Region evaluates areas along mapped (whose area is nearly 30,000 surface streams as having the best km') was published in prosects for groundwater Palmqui st and Hal 1 ( 1961 . The development . hydrogeologic maps indicate areas A report by Mull (1968) also of high, intermediate, and low dealt with the hydrogeology of probability of satisfactory well Fayette County, but the most yield and quality. Although this detailed groundwater investigation approach is the same as the one was in the Georgetown Buadrangle used by Hamilton, the two extending to North Elkhorn Creek assessments are often quite in Scott County. Mull considered different for the same area that the direction of groundwater (Hami 1ton, 1950: Palmquist and movement was controlled by the dip GROUNDWATER BASINS of the rocks and the topography, Region (United States Geological and presented his data on water Survey, 1983, is the most recent). levels in 54 we1 1s on the Other publications dealing structure contour map. primarily with other aspects of he A study of wells in the geology of the region have Centerville Quadrangle in Bourbon, included data and discussions of Fayette. and Scott counties the hydrogeology. MacBuown (1967) (Johnson, 1970; Johnson and located 16 springs and 2 wells in Thrailkill, 1973) was designed to a study of the Curdsville evaluate the relative importance Limestone Member, the basal unit of the various factors proposed by of the Lexington Limestone. He earlier workers. Based on found that some of the springs information (much of it from emerged at or near the contact of Hamilton, 1950) from 82 wells limestone beds and thin bentonites classified as adequate, sulfur, and other shale units, and that salt, or dry, nonparametric the vertical intergranular statistical methods were used to porosity and permability of the test the effect of a number of Curdsville was quite low. Another topographic, stratigraphic, or aspect of his investigation showed structural variables. A1 though that trends of sinkhole and stream apparent1y significant a1 1ignments were similar to joint relationships were found, the orientations in the Eryantsville interdependence of topographic and Quadrangle in the southern part of stratigraphic variables in an area the region. An expanded discusson of nearly horizontal beds made the of this relationship can be found results difficult to interpret. in Hine ( 1970) , who a1 so showed that joints and fracture traces Faust (1977), in a study of a (identified by tone on aerial six-county area (Bourbon, Clark, photographs) tended to be parallel Fayette, Jessamine, Scott, and as we1 1 and who located 5 we1 1s Woodford), prepared the first , in the Bryantsville Quadrangle, at potentiometric map in the region. least one of which was on sinkhole At the small scale of the map, it trend. appears to conform rather closely to topography. The map was based on data from more than 500 wells GROUNDWATER BASINS (Faust, 1977, p. 9), but the data are not shown. Faust also outlined The present day study has shown the recharge areas of a number of that the major flow of subsurface springs and we1 1s, including Royal water in those portions of the Spring, Spring Station Spring, and Inner Blue Grass Region Versailles Spring. Like earlier investigated is in at least 38 workers, he be1 ieves the yield of individual basins. The term wells is related both to "groundwater" is usually reserved topography and stratigraphy. for potable water in saturated There are also a number of voids which is beneath the Statewide reports that furnish potentiometric surface and hence specific hydrogeological at pressures greater than informaton within the region. atmospheric. Because these basins These reports include Van contain such water, they will be Convering (1962) on large springs, referred to as "groundwater Hopkins (1966b) on the elevation basins" although much of their of the fresh-saline water flow is unsaturated and above the interface, Whi tesides (1971) on potentiometric surface in what is specific capacities of we1 1s, and termed the "vadose zone." These a series of annual concepts wi11 be discursed 1ater. reports containing daily water Flow within each basin is level data in (currently) four dentritic, in that recharge from wells in the Inner Blue Grass swallets, sinkholes, and elsewhere THE INNER BLUE GRASS KARST REGION successi vel y coalesces to emerge The area of each basin (Table at a spring that drains the basin. 2) was estimated from the area A few such springs, such as outlined and ranges from less than Roaring Spring, Burgin Spring, and 0.5 km2 up to 15 kmz for Cove Spring have multiple outlets, the two largest. It should be usually within a few tens of emphasized that the areas given meters of each other; in two or are those that are believed to be more of the outlets dye was underlain by an integrated conduit detected during some traces. In no system? and that the catchment instance, however, did dye area of the spring is usual 1y much detection indicate flow between larger, since it includes areas of adjacent basins. In a few basins shallow subsurf ace or surf ace flow (Roaring Spring, Distillery outside the basin boundaries. The Spring, Slacks Spring, and areas given in this report Vaughans Spring), major flow (Table 2) are thus generally much appears at the surface at the small er than earl ier estimates bottom of deep sinks !karst (Spangler and Thrai l kill, 1981; windows), and discharge from Thrailkill and others, 1981: Spring Lake Spring feeds a surface Spangler , 1982) , which were based stream that flows into a swal let on the catchment area. These of the Lindsay Spring Basin. relationships are shown in Figure Although groundwater basins are -.7 a fundamental element of the In one location where surface hydrogeology of the region, there flow was observed between a spring has been little discussion by (Spring Lake Spring) and a swallet previous workers. Palmquist and (in the Lindsay Spring Basin), the Hall (1761, p. 14) considered length of the surface flow path groundwater in the entire Elue suggests that the two basins Grass Region (including the Inner should be identified separate1y. Blue Grass Region) to occur in small, self -contained units that, Other instances where such flow is with few exceptions, coincide with seen, are in the bottom of a deep sinkhole or blind valley, and the surf ace watersheds. Faust ( 1977, p. 12-13), outlined the recharge feature is considered a karst areas of selected points, window within the basin. including Royal, Spring Station, Groundwater basins were not and Versailles springs. He states defined for the short dye traces that such recharge areas general 1 y to Bailey and Paxton Springs coincide with surf ace drainage because of lack of evidence of the existence of deep integrated flow basins, and apparent1y based his conduits considered characteristic delineation of recharge both on of groundwater basins. topography and his potentiometric Some of the small er groundwater surface map. basins appear to underlie surface drainage basins (e.g. , Baker Cave, Basin Identification, Size, Gano Spring, and Santen Spring), while others do not (e.g., Cove and Location Spring, Elkhorn Spring, and Sharp Outlines of the 38 basins were Swall et) . At least some f 1ow drawn to enclose swallets from indicated by dye traces in all of which dye traces were made to the larger basins (5 kmz or major springs. hlthough subsurface more in area) passes beneath drainage from untraced swal lets surface divides, and the shape of within the basins as outlined most larger basins shows little probably also discharges at the correspondence to present or spring, details of basin shape are inferred former surf ace drainage largely unknown, especially for (e.g., Roaring Spring, Slacks basins identified by only a single Spring, Russel Cave, and Rurgin dye trace. Spring basins). In a few basins 36 BASIN IDENTIFICATION, SIZE, AND LOCATION

Table 2.-- Springs and Groundwater Basins in the Study Area. ------.------

SPH I NG GHOUNDWfiTER BASIN Name flagni tude Name EIrea (kma)

Bai 1ey Spring Baker Cave Spring same Big Spring same Blue Spring same Boggs Spring same

Bo~neSpring Distillery Spring Bryan Station Spring - Burgin Spring same Cougar Spring - Cornett Spring same

Cove Spring same Distillary Spring same Elkhorn Spring same Eureka Spring same Gano Spring same

Gay Sink Spring Roar i nq Spr ing Hartman Spri nq same Hol 1and Spr inq same Humane Spring same 1-75 Spring same

Jennings Spring same Lindsay Spring same McGee Sink Vaughan Spring Nance Spring same Pax ton Spring -

Pin Oak Spring same Hai lroad Spring same Roari ng Spring same Royal Sprinq same Russel 1 Cave Spring same

Santen Spring same Copper. Spring same Shawnee Hef er Spri ng - Shawnee Hun Sprinq same Silver Springs same

Slacks Spring same Slacks Cave Slacks Spring Sl oans Spring Slacks Spring Spring Lake Spring same Spring Station Spring Royal Sprinq THE INNER BLUE GRASS KARST REGION

Table 2.-- Continued.

41. Steeles Spring - same 42. Swopes Spring - Roar ing Spr ing 43. Tevi 5 Spring 5- same 44. Spring 13 4- same 45. Spring 13B 4- -

46. Vaughans Spring same 47. Versailles Spring same 48. Votah Spring same 49. Wests Spring -

Sharp Swal 1et 1 Duval Cave 1 Rnsl ey Swal 1et 1

(e. g. , Lindsay Spring an'd Vaughans approached, and the term Spring), underground flow is known "interbasin areas" wi11 be used to pass beneath perennial surface for these portions of the region. streams. Within interbasin areas, infiltratinq water from slopes and lnterbasin Areas and Basin Shape shallow sinkholes is be1ieved to Relative1y few dye traces were flow in small conduits at or just conducted in the northeast below the interface between the Woodf ord County and Walnut Hi11 bedrock and overlying regolith. areas, and further work would Flow is general1y down the probably result in the enlargement topographic slope and emerges at of the known groundwater basins small, often ephemeral, high-level and the discovery of new basins. springs. Streams fed by such While similar results would be spings generally flow on the likely near the margins of the surface but may be diverted into Mercer County area and the shallow subsurface conduits northern Fayette and southern adjacent to the stream channel for Scott counties area, intensive short distances. If and when such reconnaissance in the central a stream enters a groundwater portion of these areas (especially basin, its flow is diverted the latter) has shown that swallet underground by a swallet to emerge are much less common between the at a major spring, often several outlined basins. Furthermore, dye kilometers distant. introduced in each swallets The bottoms of most of the emerged at small springs a short major stream valleys (e.g., South distance down slope after Elkhorn Creek, North Elkhorn following shallow flow paths. Creek, Town Branch, Lower Cane Examples of such traces (none over Run, and the Salt River) appear to 500 m long or with a vertical drop lie in interbasin areas. Faust of more than 3 m) were those to (1777, p. 12 and plate 3) Paxton Spring and to Bailey described losing reaches on both Spring. North and South Elkhorn Creeks, This absence of deep, and gaging stations on these integrated, subsurface drainage creeks not uncommonly report no between basins is more marked than surface flow (U.S. Geological the simple reduction in site of Suirvey, 1983). It is likely, conduits that might be expected as theref ore, that a portion of the the divide between basins is flow of the major surface streams 38 INTERBASIN AREAS AND BASIN SHAPE

groundwater basins; the boundaries between adjacent catchment areas within an interbasin area probably conf orm close1y to surf ace divides. Although the shallow subsurface flow described above is characteristic of interbasin areas, it also occurs within the basins as out1ined. As an example, the traced swallets in the Shawnee Run Spring Basin are fed by flow from high level springs within the basin, and there appear to be extensive and numerous areas of such shallow subsurf ace f 1ow within many of the basins. An alternative way of depicting such basins would be as narrow strips adjacent to the major flow conduits, but since the location of these conduits is generally unknown, and because there is some Figure 2. Map showing relation- evidence from wells that at least ship of groundwater basins the Slacks Spring Basin is (dashed outlines) to surface developed over a considerable streams (solid lines) and sur- area, as discussed below, this was face divides (dotted lines). not done. Catchment area of spring C shown Qttempts to more closely define by dotted pattern. Although dia- the boundaries between basins and grammatic, map approximates the interbasin areas were also eastern portion of the northern compl icated by evidence that such Fayette and southern Scott coun- boundaries cannot be simply ' ties area, where A through E are depicted in two dimensions, the Silver Spring, Slacks because basin flow conduits appear Spring, Royal Spring, Vaughans to be developed beneath what Spring, and Russell Cave Spring appear to be interbasin areas in a basins, respectively. Long few cases. This situation is dashes indicate line of section illustrated by the Lindsay Spring of Figure 3. and Silver Springs basins, in is diverted into conduits through which the major f 1ow conduit swallets in the channel and passes beneath streams (fed by returns in inconspicuuus springs high-level springs) that remain in the stream bed. Such conduits entirely on the surface. are probably shallow, as are the In contrast to the conduits conduits in interbasin areas at in upland interbasin areas that higher elevations, but may be of are just beneath and roughly considerable size because of the para1lel to the land surf ace, the larger flow volumes. They are major flow conduits in groundwater probably present mainly in the basins appear to have gradients vicinity of the channel, but may similar to surf ace streams and cut across bends and meander thus are nearly horizontal and 1oops. only slightly above the level of Thus while there is only the discharging spring. Although shallow subsurface flow in the path followed by water interbasin areas, the interbasin immediately after it enters a areas form part of the catchment swallet is usually unknown, in a area of major springs draining few instances where it can be THE INNER BLUE GRASS KARST REGION obs~rvedin caves and pits it is discussed in some detai l be1ow; usual 1y steep and of ten near1y "blind valleys" which terminate vertical. Such high gradients were downstream as the entire flow of a observed as often near the margins surface stream is diverted and upstream portions of basins as underground; "pocket valleys," in the center and downstream which begin abruptly upstream at a portions. major spring: and "karst windows, " The evi dence avai 1abl e , depressions in which a major theref ore, suggests that the base underground flow emerges at the of the zone of active meteoric surface as a spring and is then water circulation is nearly flat diverted underground (the 1ength in groundwater basins (and as much of the surface flow varies from as 30 m deep beneath what appears to be a pool in the topographically high areas), rises bottom of a deep sinkhole (e.g., abruptly at basin margins, and is McGee Sink) to a stream several within a few meters of the surface hundred meters long flowing in in interbasin areas. Thus what may be described as a groundwater basins in the Inner combination of a pocket valley and Elue Grass Region are believed to a blind val ley (e. g. , the channel resemble U-shaped valleys, as below Spring Station Spring). The shown in Figure 3. flow in these landforms is major subsurface flow at or very near GROUNDWATER BASINS AND the potentiometric surface. The KARST LANDFORMS numerous sinkholes in the region The Inner Blue Grass Region is that contain a small stream fed by characterized by landforms such as a high-level spring which sinks in "sinkholes, " the distribution o+ the bottom of the sinkhole are not which was used to define the area karst windows) : and "paleoval leys" of the region, and which will be

Lurface watershed-[ 6 -5m --I km cn- cn- Basin C +I spring catchment

/

I* Basin A -I b~asinc*I Ic- asi in E +I

Figure 3. Cross section of groundwater basins and interbasin areas along line on Figure 2, showing aquifer (lined pattern) and base of zone of meteoric water circulation (lower limit of lined pattern and dashed line). Note portions of basins A and E with interbasin area characteristics (penetrated by deep flow in basin A). The relationship of basin C to the catchment area of the draining spring and the surface watershed of the stream that overlies it is also shown. Sinkholes in- dicated by S. Vertical exaggeration approximately 100X. SINKHOLE ORIGIN which appear to be normal surface conduits, solution at the base of valleys but contain no surface the bedrock will be accelerated in stream channel. Pal eoval leys the vicinity of the larger usually contain a series of conduits, and the more rapid sinkholes in their bottom, and lowering of the bedrock interface apparently formed when their nearby will cause the capture of surf ace 'stream was diverted more flow from adjacent conduits, underground at several points and hence increased bedrock along its course, forming a series solution. When the resulting of blind valleys, f 01 lowed by of the over1ying complete abandonment of surface regolith (which initially is flow except possibly during high reflected by a simple flattening discharge events. in the surface slope) is Except for some sinkholes, all sufficient to reverse the downhill five of these landforms are the slope, a topographic is result of deep circulation of formed and a type one sinkhole subsurface water, and their results. presence in an area should The existence of a topographic indicate the existence of a depression wi11 further accelerate groundwater basin, allowing the the enlargement of the conduit, location and extent of basins to since most of the water that be at least estimated from an infiltrates the surface within the examination of the topographic depression will flow through it maps. Although some correlation (although some of the flow will appears to exist, it has not been probably still be carried by possible to rely heavily on it smaller conduits). Major deepening because of sinkhole modifications and widening of the sinkhole will and the inadequacy of available probably not occur, however, until maps. Eef ore exami ninq these the conduit becomes enlarged by factors further, a discussion of solution throughout its length to the origin of sinkholes in the the degree that the water flowing region is appropriate. through it can transport particles of regolith, after which time the Sinkhole Origin iepression becomes a type two Contrary to widely held and sinkhole. The volume of regolith stated opinion, the collapse of removed may now exceed the amount the roofs of caves is not the o/ limestone dissolved, to the principal cause of sinkholes in extent that bedrock is exposed on the region (nor, for that matter, its sides or bottom. Although it in any other karst area with which seems likely that a topographic the author is f ami 1iar) . Of the depression is general 1y formed many sinkholes examined in the prior to the onset of reg01 ith region, cave roof collapse is not removal (i.e., type one precedes believed to be a major factor in type two) , this may not always be the origin of any. Rather, they the case, especially since the are produced by solution of the general downslope movement of limestone bedrock at the contact regolith on hi1lslopes wi11 tend with the overlying regolith by to fill type one depressions or water that has infiltrated from prevent them from forming. the surface, the same process that A Aype three sinkhole is formed occurs nearly everywhere in the when the conduit is large enough region and has probably been the and flow velocities are high principal agent in the lowering of enough for insoluble or otherwise the bedrock surface through time. resistent beds that tend to perch Although there will be some the conduit to be eroded through. penetration of the bedrock under a Type three sinkholes have steep or hill slope through many closely near vertical drains to depth and spaced, very small diameter their flow is integrated into the THE INNER BLUE GRASS KARST REGION dendritic system of a groudwater some cases, however, the location basin. The various types of of such drainage lines was sinkholes are shown in Fig. 4. controlled by reduced resistence Conduits draining type one and to erosion of the bedrock, due to type two sinkholes, as well as jointing or other factors, which those draining pre-type one areas would also promote more rapid (incipient sinkholes) , are usual 1y conduit enlargement. nearly horizontal, as would be Returning to the idea that expected from their being perched sinkholes are due to the collapse on resistent beds. They emerge on of cave roof 5: the growth, and nearby hillslopes or the heads of especially the deepening, of a small val leys as small , of ten type three sinkhole obvious1y is ephemeral, high-level sprinqs, highly dependent on the efficiency some of which become turbid during with which regolith and other high discharges, indicating the iebris can be removed through its sinkholes they drain have reached near vertical drain. Sinkholes their type two stage. located above conduits in the Type one and type two sinkholes under1yi ng groundwater basin are found throughout the region, system need re1ati vel y short both in groundwater basins and drains to discharge sediment into interbasin areas, and imply no the effective transport deep circulation of subsurf ace environment of the 1arger conduit , flow. Type three sinkholes, on the and are more likely than other other hand, do characterize sinkholes to deepen rapidly, groundwater basins. possibly to the point where they The tendency of sinkholes to break through into the underlying occur along former lines of conduit. A relatively minor factor surface drainage is due mainly to in this process (which is believed their development being favored by to be responsible for the formaton the increased infiltration and of karst windows in the region) subregolith flow in such areas. In may be some collapse of the roof of the underlying conduit in response to the deepening of the Incipient over1yi ng sinkhole and enlargement no depression of its drain.

bedrock Finally, it should be noted that in every instance of collapse at the surface in sinkholes known

no regolith removal to the writer, the collapse has been caused by the rapid subsidence due to transport of regolith by infiltrating water within a type two or more commonly Type Two horizontal drain a type three sinkhole, and no col1 apse of bedrock is involved. The balance between water and regolith transport through the sinkhole drain suggests that such Type Three vertical drain events should be common, but their occurrence has been greatly influenced by the practice of sinkhole filling, discussed below. Regolith collapse outside of sinkholes (i.e., not in topographic depressions) is not uncommon as well. All such col lapses the writer has examined Figure 4. Types of sinkholes were due to the failure of the 42 SINKHOLE FILLING AND MAP INADEQUACY roof of a shallow conduit that indicate the presence of a developed at or above the groundwater basin. Few of the reg01 ith-bedrock interface. streams in pocket valleys and karst windows are shown, probably because they are so short and Sinkhole Filling and Map Inadequacy hidden by vegetation and shadows. In some cases, type three Many blind valleys and sinkholes, which indicate the paleovalleys are shown as normal presence of a groundwater basin, surface valleys, especially when can be identified rather easily on the reversed slope below swallets the 1: 24,000 topographic map is gentle and short. Finally, lcontour interval 3.0 or 6.1 m) of swallets are too small to be the region. The method used is to termed landforms or to be shown determine the minimum length even by accurate maps, although necessary for the bottom of the their presence is indicated in sinkhole to be drained by a near some blind valleys. Swallets along horizontal conduit. If this length surf ace streams and in sinkholes is greater than two or three (many sinkholes do not contain hundred meters it is quite open swallets) can only be located unlikely that such a horizontal in the field. sinkhole drain exists and the sinkhole is judged to be a type GEOLOGIC AND OTHER FACTORS three. Unfortunately, the depth of INFLUENCING SUBSURFACE FLOW AND sin khol es , especi a1 1y the deeper GROUNDWATER BASIN DEVELOPMENT ones of small areas, is almost A major objective of the study always several meters greater than on which this chapter was based the depth depicted on the map by was to evaluate the degree to topograhic contours, since shadows which subsurface flow and the and dense vegetation obscure their location of groundwater basins bottoms on aeria1 photographs. delineated by dye tracing during Deep sinkholes less than 50 m this study could be explained by across are seldom shown at all on geologic and other factors. Such the topograhic maps. Many type an explanation would not on1 y three sinkholes can be identified contribute substantially to an as such only by field understanding of the nature of reconnaissance. subsurface flow in the reqion.- but A second factor hinders the would allow the prediction of flow identification of the type three directions and location of sinkholes, and hence groundwater groundwater basins in portions of basins, even after field the region where dye- tracing has reconnaissance. Deep sinkholes not been done. with steep walls provide A particular emphasis was convenient sites for rural waste placed on the relevance to disposal ; of ten the 1ong-term goal subsurface flow of the geological is to nearly fill them and render information contained on the U.S. them suitable for pasture or even Geological Survey geologic maps of row crops. This effort by farmers the area investigated ( Black, has presumably been underway for 1964, 1967: Cressman, 1964, 1967, much of the 2 centuries of 1972; Kanizay and Cressman, 1967; agriculture in the region, with Mi1 ler , 1967; MacQuown and the result that many sinkholes Dobrovolny, 1968; Pomeroy, 1968, that are actually type three now 1970: Cressman and Harber, 1970; have a shallow saucer shape more and Allingham, 19721, inasmuch as characteristic of type one or type simi 1 ar 1arge-scal e ( 1: 24,000) two. maps are available for the entire The topographic maps of the Inner Blue Grass Region. region do not accurately depict Previous hydrogeologic many of the other karst landforms investigations of the region have THE INNER BLUE GRASS KARST REGION dealt main1y with the availability Run Bed, well above the level of of subsurface water, and have major streams, and is downdip from reached varying conclusions as to its traced groundwater basin. the importance of various factors. None of the other 37 major Hamilton (1950) believed the springs draining groundwater argillaceous units in the basins in the study area indicate Lexington Limestone were the most control by stratigraphic units in important control of solution the Lexington Limestone. It would devel opment , and Mu1 1 ( 1968 seem reasonable that the few that considered them a major factor. emerge somewhat above the level of Palmquist and Hal 1 (1961) , Hopki ns major surface streams (e.g., Gano ( 1966a) , and Faust ( 1977) , on the and Steeles springs) are perched other hand, did not emphasize the on argi 1 1 aceous or otherwi se role of 1i thology, and considered resistant beds, but such beds, if topography to be the major factor. present (such as. the Macedonia Bed Mu1 1 (1968) ascribed such an at Gano Spring) are not indicated important role to the dip of the on the geologic maps or rocks that he presented his well accompanying lithologic data for the Georgetown Quadrangle descriptions, and were not on a structure contour map. observed in the field. Hami 1ton ( 1950) , Palmqui st and The control of shallow Hall (19611, Hopkins (1966a), and subsurf ace flow in interbasin especially Faust (1977) believed areas (including such areas within joints and faults played a groundwater basins) by mapped or significant role in subsurf ace unmapped argillaceous limestones f 1ow and solution devel opment . The appear to be more common. Not only previaus work utilizing infrequently, two or more traced flow paths was by Jillson high-level springs will emerge at (19451, who emphasized the the same stratigraphic level, and geomorphic development of the flow in the Sinkhole Plain paleovalleys to Royal Spring and indicated a number of such springs emerge at indirectly that downdip flow was a the top of Macedonia Bed. factor in its development There may be occasional (Jillson, 1945, p. 25-27). perching of surface streams for short distances on argi 11aceous units (e. g. , the middle reaches of Lithology of Stratigraphic Units Cane Run on the Cane Hun Red and draining groundwater basins in the the stream in the Joyland Cave study area, two are interpreted as blind valley on the Brannon being perched on argillaceous Member), but such instances are units in the Lexington Limestone. not obvious nor widespread. In one, the perching is observable Because of the general and seeming1 y clear cut; Shawnee para11 el ism between bedding and Hefer Spring in the Mercer County the overall topographic surface in area flows from a number of the areas studied, most of the hillside outlets over a distance major flow conduits and springs of 60 m along the outcrop of the are in the lower exposed units of Macedonia Bed. Although no dye the Lexington Limestone, introductions were detected at the especial 1y the Grier Limestone, spring, its groundwater basin but the stratigraphic position of probably lies to the southeast, springs emerging from this unit updip from the spring. The varies over more than 12 m, and interpretation is only slightly there is no evidence of 1ithologic , less certain for Spring Lake control. Similarly, those smaller Spring in the northern Fayette and groundwater basins that southern Scott counties area, approx imate1 y coincide with which emerges at about the surface drainages have their stratigraphic position of the Cane margins beneath surface divides BEDDING ATTITUDE that are often underlain by higher that appears to represent the argillaceous units such as the crest of the Cincinnati Qrch. Millersburg Member and the Clays Because of the problems Ferry Formati on. The numerous associated with detai led examples from both small and large structural mapping of basins which do not show this stratigraphic units that often accord with topography, however, show rapid lateral changes in indicate lithologic variations in thickness and lithology, and whose the Lexington Limestone are of esposures may be subject to little or no importance in slumping and rotation on control 1 ing the development of hi1 lslopes, the structure contours major flow conduits or the shown on the geologic maps may not location of groundwatr basins. accurately reflect local bedding Subsurface flow in major conduits attitude everywhere. If such local occurs beneath all seven of the structure is ignored and the argillaceous units mapped in the orientation of flow directions to area (Table 1). as follows (the the original dip is examined, no location of an example is in more consistent relationship is parentheses): Maced~niaBed found. In the northern Fayette and (Burgin Springs Basin), Cane Run southern Scott Counties area, Member (Royal Spring Basin) , while flow in the Royal Spring, Greendale Lenti1 (Silver Spings Slacks Spring, and Nance Spring Basi n , Mi1 1 ersburg Member basins is to the north-northwest ('Russel1 Cave Spr ing Basin) , and and down the regional dip, flow in the lower part of the Clays Ferry the adjacent Silver Springs and Format ion (Cove Springs Basin) . Lindsay Springs basin is to the southwest along regional strike. In the Mercer County area the Bedding Attitude regional dip is to the west, as is the general flow directions in The para11 el ism between bedding basins draining to the Salt River and the overall topographic Ie.g., Big Spring and Eureka surface mentioned above also Spring basins). In basins draining complicates the evaluation of the to the Dix and Kentucky Rivers importance of the dip of the rocks (e.g., Purgin Spring and Shawnee in determining flow directions in Run Spri ng basins) however, f 1ow groundwater basins. There is no , is generally to the east and hence evidence, however, of any useful up the regional dip. relationship between flow directions as indicated by dye Faults, Joints, Sinkhole Trends, traces and the dip as shown by structure contours on the geologic and Similar Features maps. Although flow in some of the & number of steeply dipping or smal ler basins is approximately vertical planar structural downdip (e.g., Versailles Spring, features, including faults, Votah Spring, Jenning Spring mineralized ,veins, and joints, are basins), in others it is nearly shown on the geologic maps of the updip (e.g., Distillery Spring, areas studied. In addition, linear Duvall Cave, Gano Spring basins). trends of sinkholes are shown by Flow directions in the larger topographic contours and others basins appear to be simi 1 arl y are visible on aerial photographs unrelated to local dip. In the iThrailkil1 and others, 1983). Lindsay Spring and Silver Springs Four of the 39 major springs basins, flow conduits cross mapped draining groundwater basins emerge anticlines and synclines at right at or within a few tens of meters angles, and in the Russell Cave of a mapped fault. In two of Spring basin the discharging these, 1-73 Spring and Nance spring and dye input points are on Spring, the dye introduction opposite 1imbs of an antic1ine points (only one for 1-75 Spring) THE INNER BLUE GRASS KARST REGION

were along the fault or an an unmapped mineralized vein in a apparent (but unmapped) extension, cave in the Shawnee Copperhead and the major flow conduit for the Spring Basin. The conduit basin is probably along. or very intersects the barite vein in near the fault. In the Shawnee Run several places at various angles Spring Basin, the spring is on the and appears to be unaffected by downthrown (about 2 m) side of a its trend. In the Silver Spring small fault that trends at nearly Basin the major flow conduit right angles to the line of flow appears to cross a mapped barite from dye introductions on the vein at about a 450 angle. upthrown side. A more complex relationship exists in the Shawnee No general relationship was Copperhead Spring Basin, where a evident between traced flow lines major flow conduit intersects an and joint directiqns, although in a few cases, as in the Silver unmapped fault and may follow it Spring Basin near the barite vein to its intersection with a mapped discussed above, the orientations fault near which the discharging of +low lines and mapped joints spring is located. are similar. It should be noted, Dye introductions were made in however, that except in a few swallets located on mapped faults places where a conduit is in three other groundwater basins. accessible and has been mapped, In the Sharp Swallet Basin, flow the only indication of the appears to follow the fault, and orientation of flow line is the the discharging spring is probably relative positions of the dye on an unmapped extension. In the input and detection points. Boggs Spring Basin, however, the flow was away from the fault (part Linear trends of sinkholes are of the same system as the 1-73 not uncommon in the Inner Blue Spring Basin Fault) at a high Grass Region. Eased on a sample, angle to the spring located some there are about 1,000 such trends distance away from its trace. identifiable on topographic maps Similarly, in the Silver Springs in the region (Thrailkill and Basin, flow from several swallets others, 1983), and hence located along a series of parallel approximately 120 in the area mapped faults is at right angles studied assuming uniform to their trend, as was flow from a distribution. Most are less than 1 swal let on the opposite side of km long, and more trend between the faults from the spring. northwest and north than in any other direction. Faust (1977, The northern Fayette and plate 2, p. 16) gave the location southern Scott counties area is of 40 such trends and stated that bounded on the southeast by the they were probabl y favorably northeast-trending Lexington Fault placed to obtain goundwater. System, a series of parallel faults with up to 150 m of mapped Aligned sinkholes are present displacement. The single dye trace along the mapped faults in the made to Bailey Spring, which lies 1-75 Spring, Boggs Spring, Sharp on the southeast (downthrown) side Swallet, Nance Spring, and Silver of a major mapped fault in the Springs basins discussed above. system, was from a swallet on the Traces from swallets on opposite northwestern side of the fault. ends of a linear trend in the The line of the trace, which was northwestern Woodford County area so short and apparently showed that the trend extends from represented such flow that the Roaring Spring to the Pin Oak no groundwater basin was defined, Spring basins. Investigations in crossed the fa~~ltat near1 y right the Royal Spring, Slacks Spring, angles. Cornett Spring, and lower Roaring I It was possible to examine the Spring basins strongly suggest relationship of a flow conduit to that the major conduit in each of 46 FAULTS, JOINTS, SINKHOLE TRENDS, AND SIMILAR FEATURES

these basins follows sub-parallel in the west (Cornett Spring Basin) 1inear sinkhole trends. to N 4S0 W in the east (Sharp Furthermore, the principal conduit Swallet Basin). The pattern does in the adjacent Sharp Swallet and not extend farther to the east, Nance Spri ng basi ns f 01 1ows mapped since the major basin is the faults (as discussed above) that Vaughans Spring Basin, whose flow are roughly parallel to these appears to follow a very well 1i near sinkhole trends. These developed 1i ne of sinkholes which relationships are shown in Table trends N 20 E. Flow in all of the 3, where the basins are listed basins is down the regional dip to from west to east. the northwest, except in the The a1 igned sinkholes , Cornett Spring Basin, where flow similarity of orientation, and is updip to the southeast. occurrence of mapped faults in two The presence of major of the basins suggests the subsurface flow conduits beneath existence of a fracture set of liner sinkhole trends was regional dimensions, with the discovered early in the study, but possibility that the fracture may the nature of the features be regularly spaced at intervals responsi ble was unknown. They were of 2 to 3 km. This hypothesis initially referred to as diaclases would suggest an additional (Thrailkill and others, in press), fracture between the Royal Springs a term which includes major and Slacks Springs basins and two ("master") joints, a set of between the Cornett Spring and closely spaced joints, or unmapped Roaring Spring basins. The first faults. interval was intensive1y Late in the study, the investigated but no groundwater opportunity arose to examine one basin was discovered in this area, of these features underground in which is on the northeastern side the major downstream conduit of of the valley of Cane Run. The the Slacks Spring Basin. The interval between the Cornett conduit, which is nearly straight, Spring and Roaring Spring basins is typically about 6 m wide and 5 has not yet been investigated. m high. It is developed in the Note that, except for the Grier Limestone Member, and the Roaring Spring Easin (which has thin, irregul arly bedded 1i mestone the least well defined sinkhole typical of this unit is exposed in trend), the orientation of the the sides of the conduit. hypothesized fractures varies Individual beds are seldom thicker rather smoothly from N 10" W than 30 cm and generally cannot be

Table 3.-- Sub-Paralled Groundwater Basins.

BASIN ORIENTQTION INTERVAL FLOW DIRECTION

Roaring Spring N 25 W N to So. Elkhorn Cr.

Cornett Spring N 10 W S to So. Elkhorn Cr.

Nance Spring N 15 W N to No. Elkhorn Cr.

Slacks Spring N 25 W N to No. Elkhorn Cr.

Royal Spring N 25 b! N to No. Elkhorn Cr.

Sharp Swal1 et N 45 W N to No. Elkhorn Cr. THE INNER BLUE GRASS KARST REGION traced laterally more than a few This interpretation may tens of meters. Visible joints can explain the rather anomalous seldom be traced more than a meter situation in the lower Vaughans or so vertical 1y, and those Spring Basin, where the path of para11 el to the conduit seldom the major conduit down flow from a extend for more than 10 meters. karst window is along a linear Over most of the 1-km trend of sinkholes, but then accessible length of the conduit, passes beneath North Elkhorn Creek the ceiling is the nearly flat to the spring on the opposite underside of an unusual 1y side. It is presumed that the continuous tabular limestone bed, conduit is developed along a a lithology more characteristic of fracture that has localized the the Tanglewood Limestone Member. sinkhole trend but is beneath a The trace of a joint, apparently resistant bed at the creek, rising little enlarged by solution, is through it on the far bank at the visible in the ceiling in many margin of the bed or at one of the places. This joint parallels the few points it is penetrated by a conduit and can be observed in solution opening. It would seem several places to be continuous likely that the spring, which is for at least 50 meters. The flat on the inside of a meander loop, cei 1ing (often several meters was once on the opposite wide) is due to collapse of weaker (southern) side of the creek, and beds up to a more resistant and that the creek channel has continuous bed, which is a common migrated laterally on the process in the nearly horizontal resistant bed. beds of the region. Topography Thus, it is beleved that a1 ignment of sinkhol es and There appears to be no localization of major conduits in consistent correl ation between the absence of faults is groundwater basins and surface controlled by the presence oi a drainage basins. Several of the joint that, unlike most of the smal ler groundwater basins (e. g. , joints in the region, is Baker Cave Spring, Humane Spring, continuous both horizontally and Gano Spring, Santan Spring, and vertically (at least 30 m in the Tevis Spring basins) appear to at one observed, judging by the depth 1east approx imatel y under1i e of the conduit beneath the surf ace drainage basins. In other surf ace). The presence of such a small basins, however (e.g., Pin joint will promote the development Oak Spring, Cove Spring, Hartman of deep sinkhole drains near the Spring , Sharp Swal 1et , and Elkhorn surface, and hence type three Spring) , subsurf ace f 1ow 1ines sinkholes (as discussed earli er) . cro5s surf ace divides. A1 1 of the &tdepth it will furnish a larger groundwater basins extend favorable path for initial conduit beneath surface divides. Examples development if it trends at a include (surface divide is in small angle to the early potential parentheses) : Big Spring Basin gradient (as wi11 be discussed (Salt River-Kentucky River), Nance below). Such conduits will more Spring Basin (North Elkhorn-South likely form in thin bedded Elkhorn creeks) , Si lver Springs 1imestones with close1y spaced Basin (North Elkhorn Creek-Cane joints, and little enlargement of fiun). In addition, in no instance the joint in massive and were the boundaries of groundwater horizontally extensive beds (such basins related to the divides of as forms the ceiling of the paleovalleys, such as the Lees conduit as described above) would Branch paleoval ley in the be expected with the exception of northeastern Woodford County area occasional near-verti cal sinkhole or the Sinkhole Plain paleovalley drains. in the Mercer County area. In contrast, the flow direction of directons or groundwater basin the shallow subsurf ace flow in boundaries that are different from interbasin areas is believed to be those now active. generally accordant with surface Prior to the Mercer County area drainage as discussed earlier. study, it was hypothesized that A1though the flow direction in the degree of groundwater basin groundwater basins appears to bear development would be less near the no consistent relationship to the margins of the region and in other details of present topography, areas where the Lexington there does seem to be a tendency Limestone has more recent1y lost for such flow to be toward the its cover of the overlying nearest major surf ace stream. In argillaceous Clays Ferry the Mercer County area, Formation. Such a relationship, groundwater basins apper to be which was discussed brief 1 y in developed on either side of a 1ine Thrailkill and others (in press), drawn midway between the Salt was not born out by the Mercer River to the west and Herrington County area study, where well Lake (Dix River) and the Kentucky developed groundwater basins River to the east. Similarly, in (e. g. , Baker Cave Spring and Cove the northern Fayette and southern Springs basins) are adjacent to Scott counties area, groundwater and even beneath outcrops of the basin flow is general 1y away from Clays Ferry Formation. a line midway between South Elkharn Creek and Town Branch on Conclusions and Utility of the southwest and North Elkhorn Geologic Maps Creek to the north and east. These The preceding anal ysi5 flow directions would correspond indicates that no single factor or to the slope of the potentiometric simple combination of factors surface of a regional aquifer appears to control the location of (which does not now exist) groundwater basins or direction of discharging along these major subsurface flow within them. The streams. best predictor of general flow direction would seem to be Geomorphology proximity to a major surface There have been easily stream, in that most of the flow interpreted changes in the in most of the basins in the areas landscape re1ated to the investigated was generally toward development of underground such streams, prabably in response drainage. The upper portions of a to a potentiometric gradient in number of surf ace val 1eys have existence early in the development been converted into blind valleys of the subsurface flow systems. and in a few cases paleovalleys Groundwater basins will be have been created by the diversion found beneath deep sinkholes, underground of essenti a1 1y a1 1 bli nd val 1eys, and paleoval 1eys, surface drainage. Similarly, in but the lack of such landforms several o+ the caves of the region does not necessarily indicate the passages that are not now carrying presence of interbasin areas. subsurface flow are found a few Where the trend of aligned deep meters above the active flow sinkholes does not deviate from conduits, and there are high-level the direction of the early openings near a few of the major potentiometric gradient by a large springs (e.g., Roaring Spring, angle, it is likely that major Lindsay Spring) that probably basin conduits are developed represent abandoned conduits beneath such an alignment. (although most of these are A11 of the above features are utilized during high flow). None shown, with varying degrees of of these higher level conduits, accuracy, on the topographic maps however, indicate earlier flow of the region. The principal THE INNER BLUE GRASS KARST REGION information presented on geologic that of areas underlain by maps, the areal extent and granular material are so lithologic nature of stratigraphic substantial that virtual1 y the units in' the Lexington Limestone, only feature the two systems have is of little or no utility in in common is the presence, flow, locating the boundries of and flow and avai 1abi 1i ty of water beneath directions within groundwater the surface to wells. Because a basins, nor does bedding attitude fundamental starting point for the as shown by structure contours description of the hydrogeology of provide usef ul information. About granular , and the the only features delineated on overlying vadose and regolith geologic maps (and not on zones, is that the type of flow is topographic maps) that may be of such that Darcy's Law is followed, interest are faults along which an examination of the types of aligned sinkholes are not present, subsurface flow in the Inner Flue although no conduits were shown to Grass Karst Region is appropriate. follow such faults in the area studied. It is possible that there Types of Flow is a slight tendency for basins in Subsurface flow in an area which the flow is down the underlain by granular material is regional dip to be enlarged largely through pores of such relative of those in which flow is small diameter that the flow updip, but no real evidence of velocity is linearly related to this was seen during the study. the potential gradient by the hydraulic conductivity, a NATURE AND DEVELOPMENT OF THE relationship described by Darcy's HYDROGEOLOGIC SYSTEM Law. In addition, flow in small planar fractures (e-g., joints and The following discussion may be bedding surf ace) wi11 a1 so obey premature, inasmuch as no studies this relationship if the width of in the region of important topics the fracture is sufficiently such as water budget or carbonate small. The term "capillary size" geochemistry have yet been will be used here, although compl eted. The re1at i onshi ps capillary effects are pertinent established during the present only in unsaturated flow. If the study, however, provide a pores (and fractures) are not framework for an explanation of saturated with water, the flow the nature and development of the will be termed "unsaturated hydrogeology of the system that is intergranular flow" (and the sufficiently different from the degree of saturation is an added views of earlier workers to parameter in flow relationships): justify its presentation. otherwise the flow will be termed The ideas that will be "saturated intergranular flow. " presented are based on arguments Although other types of flow may that are rather highly deductive. occur, as in large soil fractures The only portions of the and in areas of high potential subsurface system that can be grad ient near pumpi ng we1 1s , they directly observed in any detail may usual 1y be safely neglected in are conduits that are large enough describing the hydrogeologic to enter and are not completely system. The body of saturated water f i1 led. Although consistent granular material at depth in with observations made during the which saturated intergranular flow study, the properties of, and occurs, and in which the water processes occurring within, the pressure is greater than one smaller conduits must maini y be atmosphere, is considered the deduced from physical principles. "aquifer" (and its contents The differences between the "groundwater") if its hydraulic hydrogeology of the region and conductivity is high enough for TYPES OF FLOW water to be yielded to wells. scale of a few weeks to a few Above the "potentiometric surface" years in considering the water (termed the "water table" if the budget of the region. As will be aquifer i5 unconfined) , at which discussed , however. such f 1ow i5 the pressure is atmospheric, most important on a long (i.e., of the flow is unsaturated geological time scale in intergranular flow although a understanding the development of region of saturated intergranular the hydrogeologic system of the flow (lower portion of the region. Note that the two types of capillay fringe) is usually intergranular flow include flow present just above the along narrow fractures, as well as potentiometric surface in the that between grains. "vadose zone" and, locally and The other four types of flow temporarily, in portions of the are in conduits, which are regolith as a result of high solutionall y enlarged openings recharge. larger than the capillary size In contrast, subsurface openings so far discussed. meteoric water in the Inner Flue Although many conduits are tubes Grass Region is transported by six rather than regular cross sections different types of flow, all of that change little along the which are significant in length of the conduit, the term describing the nature and will also be applied to all large development of the hydrogeol ogi c openings in the rock regardless of system. In the regolith, flow is their shape. similar to that in the regolith "Pipe flow" occurs when the over1ying granular material, and condui t is completely f i11 ed with water is transported largely by water and (since there are no unsaturated intergranular flow, capillary effects and the venturi with areas of saturated effect of high velocities is intergranular flow beneath negligible) the pressure is and surface streams as well as greater than atmospheric. The elsewhere f 01 lowing heavy rains or other types of conduit flow are snow melt. Unlike many areas of unsaturated (i.e. , the conduit granular rocks with appreciable contains both water and air). In hydraulic conductivity, however. a "bedrock channel flow", flow is on zone of saturated intergranular bedrock beneath a free surface, flow is often present above the and hence the width, depth, and regolith-bedrock interface due to gradient are fixed for a given the very low hydraulic discharge except for solution and conductivity of the bedrock if no abrasion of the bedrock on a long conduits are developed. In time scale. "Gravity flow" differs addition, one or more of the four from bedrock channel flow in types of conduit f 1ows discussed having a very high gradient, below may occur in the regolith lacking a well defined (especially its lower part) in cross-sectional area, and having conduits excavated by piping and poorly defined contact (or none in other non-solution processes. the case of water f a1 1ing free) Flow in the bedrock outside of with the bedrock, which precludes conduits will be by saturated the application of open channel intergranular flow as we1 1. flow relationships (e.g., elthough this is overwhelming1y Chezy-Manning) used for other the largest region in the types of unsaturated conduit subsurface, intergranular flows. Finally, "equilibrium hydraulic conductivities in the channel flow" is similar to bedrock are so low that this flow bedrock channel flow (and is is of no interest on a short time describable by open channel flow scale as a source of water to relationships) except that the wells nor on an intermediate time bottom and sides of the channel THE INNER BLUE GRASS KARST REGION 5 1

are large1y on sediment, main1y filthough other types of transported regolith and bedrock subsurface flow may occur in the fragments, and its width, depth, region. such as in saturated or and gradients on a long and unsaturated conduits in areas of possibly intermediate time scale ponding or in saturated conduits are determined by an equilibrium partly fiiled with sediment, it between water and sediment may be assumed, at least transport. Such flow has been initially, that such flow may ex tensi vel y discussed (under a adequate1y be described as one of variety of names) by many authors the types described above. The for surface streams (e.g., Leopold proprties of the six types of flow and others, 1964; Hammer and considered are summarized in Table MacKichan, 1981 ) . 4.

Table 4.-- Types of Subsurface Flow in the Inner Blue Grass Region.

PREDOMI- POTENTIAL PRESSURE TANT VELOCITY TYPE OF SATURATED OR TYPE OF REL. TO FLOW RELATION- FLOW UNSATURATED OPEN ING CITM. MODE SHIPS

Saturated Saturated Capillary Greater 1ami nar Darcy intergran- (occ. about ular flow equal or 1ess)

Unsatur- Unsaturated Capi 11ary Less 1 aminar Darcy ated (modified)

intergran- ular +low

Gravity Unsaturated Conduit About turbu- Gravita- Flow equal lent tional accel era- tion vertical film, etc.

Pipe Saturated Conduit Greater turbu- Turbulent Flow lent pipe flow

Bedrock Unsaturated Conduit about turbu- , Chezy- channel equal 1ent Manni ng , flow etc.

Equi 1i - Unsaturated Conduit about turbu- Chezy- brium equal lent Manni ng , channel Leopol d , flow etc. THE NON-METEORIC SYSTEM

The Non-Meteoric System Conduit Initiation

Before proceeding further with Virtually by definition, the a discussion of the nature and flow in bedrock to conduit development of the subsurf ace development is by saturated meteoric water flow system, some interqranular flow, and such flow 1s now occurring in bedrock where mention of what will be termed the conduits are not present. An non-meteoric system is in order. examination of the transition from As discussed earlier in the intergranular to conduit flow section on water supply, many would thus seem to be an essential wells drilled in the region part of the development of the encounter water of unsatisfactory flow system, but as the following quality, in some cases at depths will show, no very satisfactory of less than 25 m. This water is conclusion can be reached variously rharacterized as regarding this phase of the containing sulfur, salt, iron, hydrogeolgic history of the etc., and may be present in region. appreciable quantities in some The principal mechanism wells. resoonsi ble for the initiation of Although little is known of conduits is solution of the this subsurface water, several mineral calcite, and principal observations can be made. First, constituent of limestone, and at least some of the water is in although various attempts have conduits (and presumably by pipe been made to quantify the flow at these depths), inasmuch as relationships between solution and the intergranular hydraulic flow (e.g., White, 1977), much conductivity is too low to work remains in this area. It is transmit the amounts of water that evident, however, that conversion have been encountered. Second, the of an intergranular flow path to a chemistry of the water indicates conduit flow path requires the that it is isolated from the passage of large amounts of water meteoric water system. Third, the simply to remove the solution absence of such water in many deep products, regardless of the dry holes and underground details of the soluton kinetics or suggest tht this system does not degree of chemical undersaturaton completely permeate the bedrock. of the water as it enters the flow Fourth, the apparent difference in path. Assuming a high and constant chemistry of this water suggest partial pressure, that it may be in small, no dissolved calcite in the water relatively isol ated bodies, and as it enters the flow path, and that a continuous system does not complete saturation with respect exist. Final1 y, the fact that some to calcite as it leaves the flow wells that initially yield water path (a11 unreal i5ti call y generous of unsatisfactory qua1 ity later specifications), a volume of water produce meteoric water, suggests at 1east 1,000 times the volume of that pressure communication the initial conduit (neglecting between the non-meteoric and the volume of the intergranular meteoric systems may exist, and f 1ow path ' is needed dcrri ng the continued pumping of the former period of intergranular flow. allows the latter to invade the Assumi ng a potential gradient conduits and flush them out. of 0.01 (based on the region's A1 ternativel y, these cases may be topography) , a flow path length of explained by the well initially C km, and a minimum time for water producing from both systems which to traverse the flow path of 10 then exhausts the non-meteoric years !thus providing the above system, supporting the suggestion volume in 10,000 years), an that these are actually a series application of Darcy's Law yields of isolated systems. a minimum hydraulic conductivity THE INNER BLUE GRASS KARST REGION along the flow line of a little a fracture that has not been more than 10-s m/s. solutional ly widened, a hydraulic Intergranular hydraul ic conductivity of about 10-l1 conductivities of the limestones m/s is obtained using methods and thin shales of the Lexington described in Freeze and Cherry Limestones are 1ow. MacQuown (1979, p. 74), and the void ratio (1967, p. 68), presents a (assuming all fractures are determination equivalent to about parallel to flow) is about 2.5 x lo-' m/s for a specimen of lW3. Even if no path the Curdsville Member, which is lengthening due to tortuousi ty is lithologically similar to the considered, a flow velocity within Tanglewood Limestone Member. a fracture of 4 x lV9 Freeze and Cherry (1979, p. 29) results, one and a half orders of indicate that a hydraulic magnitude less than that of an conductivity of la-' m/s is intergranuar path. about the lower limit for Although this admittedly crude limestone, and hence this probably analysis suggests the represents intergranular, as intergranular flow paths should be opposed to fracture, hydraulic favored over fracture flow paths conduct ivi ty. during the pre-conduit flow stage, The actual flow velocity along the reverse is probably true, a flow path will be inversely since small conduits observed in related to the bulk velocity outcrop are usually, but not (suggested by the hydraulic invariably, localized along a conductivity) by the void ratio, joint or bedding surface. Thus, assuming the flow path is there may be errors and straight. A void ratio of inconsistencies in the la-=, and a degree of assumptions, most notably in the tortuousness of the flow path such specification of fracture width. that it is 10 times the straight Since hydraulic conductivity along 1ine distance, yields a flow a fracture is directly related to velocity of lW7 m/s, two the third power of the fracture orders of magnitude too low for width !Freeze and Cherry, 1979, p. conduit initiation under the 741, if the width is 1 mm conditions assumed. (lo-=) rather than 0.1 mm. Because the Lexington Limestone the hydraulic conductivity is is thin-bedded and the individual increased by 3 orders of beds are jointed, pre-conduit flow magnitude, favoring fracture paths a1 ong bedding and joint surf ace, over intergranular paths. Such a which wi11 collective1y be called width for non-solutional 1y widened "+ractures," would seem likely. fractures at depth seems too great Such flow in a system of narrow (0.1 mm is probably too generous), fractures (assuming certain but it is likely that some conditions of their solutional widening (and even interconnection and spacing are conduit development) has occurred met) will obey Darcy's Law and is in at least some fractures prior here considered saturated to the initial entry of meteoric intergranular flow, even though water. Openings large enough to the flow paths are not between transmit the non-meteoric system grains. discussed earl ier are certain1y MacBuown (1967, p. 47) found present in some places in the the average spacing of bedding rock. surfaces to be 0.05 m and the The apparent near-comparable average joint spacing to be 0.24 rn efficiency of intergranular paths in the Curdsville Member, which suggests that pre-conduit flow yields a value of 24.2 along such paths cannot be fractures/mz. Assuming a ignored, however. If a steep width of 0.1 mm (lob4 m) +or potentia1 gradient were present at STAGES OF CONDUIT GROWTH an angle to bedding where no familiar second type (branching joints were present, enlargement upf low). bf intergranblar paths para1lel to the gradient would be expected. Stages of Conduit Growth Such paths would probably even Further solutional (and cross shale interbeds up to abrasion) enlargement of the several mi11 imeters thick (which conduits and integration of the probably includes most such conduit system has led to the interbeds in the Lexington present hydrogeologic System o+ Limestone) , inasmuch as the shales the region. During this generally contain more than 50 enlargement and integration, percent calcite (and dolomite) and individual conduits have passed less than 25 percent clay minerals through a number of significant (Fisher, 1968, p. 780) , and hence stages. The transition to the even their vertical hydraulic first stage occurs when the conductivity may be comparable to cross-sectional area of a conduit the hydraulic conductivity of the becomes sufficiently large, and limestones. Conduit development in the flow velocities (due to such shales would be inhibited by integration of the conduit system) the accumulation of insoluble sufficiently high, for the flow to residue, however. become pipe flow, and hence'no longer described by Darcy's Law. Ewers and Quinlan (1981) have Prior to this transition, the flow presented the most persuasive would be saturated (usual 1 y) explanation for the initial intergranular flow, even though it development of conduits from was in the embryo conduits saturated intergranular flow a1 ong described in the preceding a fracture. Ewer's (1981) section. Because both the plan and experiments (utilizing salt and cross-section of the conduits are, plaster) indicate conduit probably quite irregular, the development begins at the input transition to the first stage point and extends down the flow as probably occurs well before the a complexly branching dendritic flow becomes turbulent. pattern of small conduits. Because potential loss in the conduits, is The transition to the second much less than in the stage occurs when conduit size intergranular flow region, the throughout its length is great steepest potential gradient is enough for sediment (both reg01i th between the outlet and the end of ad the insoluble residue from the the conduit nearest the outlet soluti onal en1argement of the resulting in increased flow and conduits) to be transported accelerated conduit growth a1ong through te system. The third stage this line. Once the first conduit is reached when the size of the reaches the outlet, potential conduit and the flow velocities falls in all the conduits and flow are sufficiently high for conduits within the growth of the other on bedding surf aces above thin conduits in the dendritic pattern shales or otherwise resistant beds virtual 1 y ceases. If dendritic to erode through to the underlying patterns of conduits are growing less resistant limestone. The from other input points, a steep conduit size and flow velocity potential gradient develops in the necessary is obviously a function intergranular flow region between of the extent, thickness, and these conduits and those of the degree of resistance of the pattern that first reached the under 1yi ng bed. outlet, causing conduits in the It seems unlikely that pattern that first reached the significant sediment transport can outlet. Thus, the first type of occur unless the f 1ow is dendritic pattern (branching turbulent, and conduits that are downflow) is converted to a more able to erode shales (probably THE INNER BLUE GRASS KARST REGION mainly by solution, inasmuch as dye tracing has indicated that the the "shales" are dominantly subsurface conduit system appears carbonates, as discussed earlier) to be deep, extensive, and well must be able to transport the integrated, while there is no insoluble residue out of the evidence that the subsurface conduit. Thus, the three stages conduit system in interbasin areas would seem to be sequential. There has any of these characteristics. is another transition that occurs In groundwater basins, at least at some point during the the major f 1ow of meteoric water enlargement of a conduit and infiltrating the surface descends integration of the system whose steeply through stage three position in the sequence may vary, conduits from stream swallets or a1 though it probably occurs most as type three sinkhole drains. o+ten during the second stage. In two of the qroundwater This transition occurs when the basins identified (Shawnee Hef er size of the conduits and and Spring Lake Spring basins), integration of the system reaches the major basin conduits are the point where the amount of believed to be perched on a water being supplied to the resistant bed, and thus have not conduit is insufficient to fill reached the third stage of it, at least during times of low development relative to this bed recharge, and the flow becomes (although third stage conduits are unsaturated, either bedrock probably developed through thinner channel flow, if the gradient is resistant beds above it). low, gravity flow, if the gradient In the remaining 36 groundwater is high (most common in a third basins, flow within them appears stage conduit) or equi 1 brium , i to be in large, nearly horizontal channel f 1ow in larger and deeper conduits, whose elevation is conduits. unrelated to lithology. Where Where the conduit serves as a major conduits can be entered and sinkhole drain, this examined, they consist of open classification corresponds to the passages traversed by a stream classification of sinkhole types f 1owing over sediment , with outlined earlier, in that accessibility terminating both incipient and type one sinkholes upstream and downstream when the are drained by first stage conduit becomes complete1 y filled conduits, type two sinkholes by with water. The nearly horizontal second stage conduits, and type gradient of these major conduits three sinkholes by third stage is believed to be controlled by conduits. the equilibrium flow occurring in As stated earlier, geochemical the unsaturated portions of the studies of the ability of major conduits. recharging meteoric water to PIS discussed earlier, the width accomplish the conduit enlargement of the zone of near horizontal have not yet been completed in the flow at depth in underground region. A considerable body of basins is uncertain. Although literature exists on this question potentiometric surface elevations based on studies in other areas, in the middle Slacks Spring Basin however (e.g., Thrailkill and suggest that it may be extensive, Robl, 1981), and it is believed other evidence would seem to that this model of conduit indicate that conduit development initiation and development is between major flow lines within consistent with the geochemistry. the basin is minor or absent, and that the basin flow is largely Groundwater Basins, lnterbasin through a single conduit or, in a Areas, and the Aquifer few cases, conduits parallel to Groundwater basins have been and very near the major conduit. identified as areas within which Such evi dence incl udes we1 1 data GROUNDWATER BASINS, INTERBASIN AREAS, AND THE AQUIFER

+,om the lower Slacks Spring Basin of rock that contains water that and other basins in the Georgetown is available to wells in useful Quadrangle, the fact that most of quantities and that is under a the springs either have a single pressure greater than atmospheric. outlet or multiple outlets very In addition, the water should be close to each other, and the fact of usable quality. The term has that impoundment of springs has been avoided so far in this report not led to their abandonment and a because the nature of the major diversion of flow as the subsurface flow system in the potential is increased. region is so different from that Subsurface flow within the in granular materials that the groundwatr basins (neglecting the term is essentially meaningless saturated and unsatrated unless carefully characterized. intergranular flow in the regolith Similarly, since the term and saturated intergranular flow "groundwater" is best reserved for in the bedrock outside of water in the aquifer, the term conduits) is thus different in "subsurf ace water " has been different parts of the basin. emp 1oyed . Water entering the basin from In the Inner Flue Grass Region, stream swallets and type three therefore, the aquifer consists sinkhole drains, initially only of rock in which conduits are descends steeply by gravity flow developed (since intergranular and short reaches of bedrock flow does not satisfy the yield channel flow to the floor of the criterion) that contain meteoric basin. It then is transported to water (the non-meteoric system the discharging spring mainly by fails the quality criterion) at equilibrium channel flow and pipe greater than 1 atmosphere flow, although reaches of low pressure. Because shallow bedrock gradient bedrock channel flow channel f 1ow and equi 1i bri um several hundred meters long have channel flow, as well as gravity been observed in the upstream flow, are at atmospheric pressure, portion of small er conduits. only rock with conduits with pipe Although it is rather easy to flow and the deeper water filled explain the near horizontal flow portions of larger conduits in in the groundwater basins as being which bedrock channel flow and due to equilibrium channel flow in equilibrium channel flow occurs at least major portions of the are included. larger conduits, it should be Within groundwater basins, the noted that other, and unknown, potentiometric surf ace is factors promoting this horizontal represented by the water surface flow may be operating. Fy its very in the larger conduits in which nature equilibrium channel flow eqilibrium channel flow is requires that large amounts of occurring. Adjacent conduits are sediment are being transported in completely water filled if they the subsurf ace. While this is are below this level, with the certainly true in the Inner Flue water pressure determined by the Grass Region, it may not be in depth below the potentiometric other karst areas where surface. Flow in other conduits near-horizontal flow a1 so occurs. that are partly above this level This equi 1i brium f 1ow explanation will be main1y by bedrock channel is not, therefore, necessarily a flow, with equilibrium channel general explanation of the causes flow in those carrying large of shallow versus deep phreatic amounts of sediment from the flow that has been extensively surface. Well data from the middle debated in the literature (e.g., Slacks Spring Basin shows that at Thrai 1ki 11, 1968). least in one basin the In hydrogeologic systems, an communication between these aquifer is considerd to be a body various conduits is sufficient to THE INNER BLUE GRASS KARST REGION produce the expected near1y flat practice of filling sinkholes potentiometric surface over a wide mentioned earlier has probably area . decreased subsurf ace flow, since It should be noted that fairly precipitation that formerly high gradient bedrock channel flow entered the subsurface rapidly occurs in many places and at many through swallets in deep sinkholes elevations in the groundwater is not retained in the regolith basins. Since the gradient is (and occasional 1y in ponds high, the f-low is rapid and established in sinkholes) and shallow. This water was excluded evapotranspired. On the other from the aquifer in the above hand, surface runoff into small definition because it is streams and into swallets that essentially at atmospheric divert their flow underground, has pressure and, since it is unlikely been increased by land clearing that the surface of such flows is and urbanization. Although the net reflected in the surface of nearby effect (to either increase or unsaturated flows or the pressure decrease recharge) may have been in pipe flow conduits; it is substantial, it cannot be meaningless as a potentiometric evaluated with the present data. surf ace. Because of the high hydraul ic In the smaller conduits in the conductivity and low specific interbasin areas, pipe flow and storage of the aquifer, however, occasional 1y large channel flows such changes in recharge rate have may be encountered near the a small effect relative to what surface, and a consistent would be expected in a granular potentiometric surface may be aqui f er . definable over a small area. Along Human activities have also major streams, 1arger f 1ows modified the flow in conduits by beneath a more continuous causing subsurface sedimentation. potentiometric surface at or just The impoundment of major spr ings above the stream level would be such as Russell Cave and Royal expected. The margins of Spring has apparent1y produced groundwater basins in extensive deposition in the topographically high areas are downstream portion of the main probably so steep that no aquifer conduit, and it is likely that the exists. series of low dams that have been Thus the Inner Blue Grass constructed on North and South aquifer is discontinuous on two Elkhorn creeks has had a similar scales. Since it exists only where effect on some of the springs conduits are developed, it can be flowing into these streams. In tapped by only a fraction of the addition, there are extensive wells that are drilled. In fills of transported regolith in addition, since it can be defined several of the accessible conduits only when pipe flow and low in the region. In some cases these gradient channel flow are are in upper level conduits occurring, it may be characterized (mainly sinkhole drains) in which as being estensi ve in groundwater" the water transport is by bedrock basins and along major sur+ace channel flow and gravity flow. streams, discontinuous and local Although some sediment would be in topographically high portions expected to be transported through of interbasin areas, and may be such conduits (and equilibrium absent at basin boundaries. channel flow might develop locally), the observed fill is far Influence of Human Activities in excess of the amount expected Some mention should be made of and does not appear to be the effects of underground flow in transported by even the highest the region as a result of human recharge events. Similarly, the activities. The widespread accessible portions of the major REFERENCES CITED conduit in the Slacks Spring Basin Byrd, F. E. , 1981, The effects of (whose spring is not impounded) two water tracing agents on contain large amounts of passive cotton dye detectors: transported regolith on either Lexington, University of Ken- side of the active equi 1i brium tucky, M.S. Thesis, 51 p. channel flow, and dates scratched into the fill indicated that much Byrd, P. E. , and Thrai 1ki 11, John, of it is not inundated, or 1983, Response of fabric detec- transported during high flows in tors to optical brightener in the conduit. It is believed, determining underground f 1ow therefore, that much of this connections (Abs. ) : Geological excess sediment may have been Society of America, Abstracts introduced into the subsurface as with Programs, v. 15, p. 96. a result of initial land clearing operations, probably in the early Cressman, E. R. , 198 1, Geol ogy of part of the 19th century. the Tyrone Buadrangle, Kentucky: Finally, it may be noted that U. S. Geological Survey Geologic underground basins exist within Quadrangle Map Gl2-303. parts of the city of Lexington, as evidenced by the presence of major Cressman, E. R., 1967, Geologic springs, deep sinkholes, karst map of the Georgetown Quadran- windows, and blind valleys. No dye gle, Scott and Fayette counties, tracing has yet been attempted Kentucky: U. S. Geological Sur- within this heavily urbanized vey Geologic Buadrangle Map area, however, due to the GB-605. difficulty of clearly distinguishing natural, subsurf ace Cressman, E. R, 1972, Geologic flow from that in storm drains. map of the Danville Buadrangle, Mercer and Boyle counties, Ken- tucky: U. S. Geological Survey REFERENCES CITED Geol og ic Buadr ang 1e Map GQ-985.

Allingham, J. W., 1972, Geologic Cressman, E. R., 1973, Lithostra- map of the Harrodsburg Buadran- tigraphy and depositional envi r- gle, Mercer and Woodford coun- onments of the Lexington Lime- ties, Kentucky: U. S. Geological stone (Ordovician) of central Survey Geologic Quadrangle Map Kentucky. U. S. Geological GQ- 1020. Survey Frof essi onal Paper 768, 61 p. Black, D. F. B., 1964, Geology of the Versailles Buadrangle, Ken- Cressman, E. H., and Harber. S. tucky: U. S. Geological Survey V., 1970, Geologic map of the Geol ogi c Quadrangle Map 68-325. Wilmore Quadrangle, central Ken- tucky: U. S. Geological Survey Black, D. F. B., 1967, Geologic Geol ogi c Quadrangl e Map 68-847. map of the Coletown Quadrangle, east-central Kentucky: U. S. Ewers, R. O., 1981, The develop- Geological Survey Geologic Quad- ment of limestone cave systems rangle Map GQ-644. in the dimensions of length and breadth: McMaster Univ., Ph.D. Black, D. F. P., Johnson, R. W., Dissertation. Jr. , and Kel ler, G. R. , 1977, Fault systems of the 38th Par- Ewers, E. O., and Quinlan, J. F., allel lineament in central Ken- 1981, Cavern porosity develop- tucky and their relationship to ment in limestones: a low dip other tectonic features in the model from Mammoth Cave, Ken- area (fibs. ) : Geological Society tucky: Eighth International Con- of America, fibstracts with Pro- gress of Speleology Proceedings, grams, v. 9, no. 5, p. 576. p. 727-731. THE INNER BLUE GRASS KARST REGION 59

Faust, R. J., 1977, Groundwater Geological Survey, ser. 9, Eul- resources in the Lexington, letin 5, 66 p. Kentucky, area: U. S. Geological Survey-Water Resources Investi- Hammer, M. J . , and MacKi chan , gations 76-113, 24 p. K. A. , 1981, Hydrolagy and quality of water resources: New Fisher, I. S. , 1968, Interrelation York, Wiley, 486 p. of mineralogy and texture within an Ordovician Lexington Lime- Hendrickson, G. E., and Krieger, stone section in central Ken- R. A., 1964, Geochemistry of tucky: Journal o+ Sedimentary natural waters of the Blue Grass Petrol ogy , v. 38, p . 775-784. Region, Kentucky: U. S. Geolog- i cal Survey Water-Suppl y Paper Freeze, R. A., and Cherry, J. A.. 1700, 135 p. 1979, Groundwater: Engl ewood Cliffs, New Jersey, Prentice- Hine, G. T., 1970, Relation of Hal 1 , 604p. fracture traces, joints, and groundwater occurrence in the Hal 1, F. R. , and Palmquist, W. N. , area of the Bryantsville Ruad- Jr., 1960a, kvailability of rangle, central Kentucky: Ken- groundwater in Bath, Fleming, tucky Geological Survey, ser. and Montgomery counties, Ken- 10, Thesis Series 3, 27 p. tucky: U. S. Geological Survey Hydrologic Investigation At1 as Hopkins H. T., 1966a, Water re- HA-18. sources of the Fayette County area, Kentucky, in application Hall, F. R. , and Palmquist, W. N. , to land-use planning and engi- 1960b, Availability of ground- neering in Lexington and Fayette water in Clark, Estill , Madison, counties, Kentucky: U. S. Geo- and Powell counties, Kentucky: logical Survey Open-Fi le Report, U. S. Geological Survey Hydro- p. 1-17. logic Investigation Atlas HA-19. Hopkins, H. T., 1966b. Fresh- Hall, F. R. , and Palm~t~ist,W. N. , saline water interface map of 1960c, Avai 1abi l ity of ground- Kentucky: Kentucky Geological water in Carroll, Gallatin, Survey, ser. 10, Scale 1 in. =

Henry, Owen, and Trimble coun- f 8 mi. ties, Kentucky: U. S. Geological , William, and Thrailkill, Survey Hydrologic Investigation John, 1993, The lack of influ- Atlas HA-23. ence of 1i tho1 ogy and structure on the development of a karst Hal 1, F. R. , and Palmquist, W. N. , aquifer (Abs): Geological So- 1960d, Avai 1 abi 1 i ty of ground- ciety of America, Abstracts with water in Anderson, Franklin, Programs, v. 15, p. 252. She1 by, and Woodf ord counties, Kentucky: U. S. Geological Sur- Jillson, W. R., 1945, Geology of vey, Hydrologic Investigation Roaring Spring: Frankfort, Ken- Atlas HA-24. tucky, Roberts Printing Co., 44 P Harni 1 ton, D. K. , 1948, Some sol u- . tion features of the limestone Jillson, 1961, Erosion cycles in near Lexington. Kentucky: Eco- central Kentucky: Frankfort, nomic Geology, v. 43, p. 39-52. Kentucky, Roberts Printing Co., 12 p. Hamilton, D. K., 1950, Areas and principles of groundwater occur- Johnson, J. T., 1970, Nonparame- rence in the Inner Blue Grass tric analysis of variables in- Karst Region, Kentucky: Kentucky f 1 uencing 1 i mestone groundwater REFERENCES CITED

occurrence: Lexington, Universi- ton and Fayette County Planning ty of Kentucky, M. S. Thesis, Commission, 24 p. 41 p. Palmquist, W. N. , Jr. , and Hall, Johnson, J. T. , and Thrai 1 ki 11 , F. R., 1960a, Availability of John, 1973, Variables affecting groundwater in Bracken, Harri- well success in a Kentucky lime- son, Mason, Nicholas, and stone aqui fer: Journal of Hy- Robertson counties, Kentucky: U. drology, v. 20, p. ~27-333.- S. Geological Survey Hydrol ogi c- Investigation Atlas HA-16. Kanizay, S. P., and Cressman, E. R., 1967, Geologic map of the Palmquist, W. N. , Jr. , and Hall, Centervi 1 le Quadrangle, central F. H., 1940b, Availability of Kentucky: U. S. Geological Sur- groundwater in Eoyle, Gerrard, vey Geologic Buadrangl e Map GB- Lincoln , and Mercer counties, 654. Kentucky: U. S. Geological Sur- vey Hydrol ogic Investigation Leopold, L. B. , Wolman, M. G., and Atlas HA-20. Miller, J. P., 1964, Fluvial processes in geomorphol agy: San Palmquist, W. N., Jr., and Hall, Francisco, Freeman, 522 p. F. R., 1960c, Availability of groundwater in Bourbon, Fayette, MacBuown, W. C., Jr., 1967, FacL Jessamine, and Scott counties, tors controlling porosity and Kentucky: U. S. Geological Sur- permeability in the Curdsville vey Hydrologic Investigation Member of the Lexington Lime- At1 as HA-25. stone: University of Kentucky Water Resources Institute, Re- Falmquist, W. N., Jr. , and Hal 1, search Report 7, 80 p. F. R. , 1961, Reconnaissance of groundwater resources in the MacQuown, W. C. , and Dobrovolney, Blue Grass Region, Kentucky: E., 1968, Geologic map of the U. S. Geological Survey Water- Lexington East Quadrangl e, Fay- Supply Paper 1533, 39 p. ette and Bourbon counties, Ken- tucky. U. S. Geological Survey Fomeroy, J. S., 1968, Geologic map Geologic Buadrangle Map GB-683. of the Frankfort East Buadran- gle, Franklin and Woodford coun- Matson, G. C., 1909, Water re- ties, Kentucky: U. S. Geological sources o+ the Blue Grass Survey Geologic Quadrangle Map Region, Kentucky: U. S. Geolog- GQ-707. i cal Survey Wat~r-S~~pply Faper 233, 223 p. Fomeroy, J. S., 1970, Geologic map of the Midway Quadrangle, cen- McCann, M. R., 1978, Hydrogeology tral Kentucky: V. S. Geological of northeast Woodford County, Survey Geol ogi c Quadrangl e Map Kentucky: Lexington, University GQ-856. of Kentucky, M.S. Thesis, 103 p.

Miller, R. D., 1957, Geologic map Quinlan, J. F., 1977, New fluores- of the Lex i ngton West Ouadran- cent direct dye suitable for gle, Fayette and Scott counties, tracing groundwater and detec- Kentucky: U. S. Geological Sur- tion: Third International Sym- vey Geol ogi c Quadrang1 e Map GB- posium of Underground Water 600. Tracing, v. 2, p. 257-262.

Mull, D. S., 1958, The hydrology Guinlan, J. F., and Ewers, R. O., of the Lexington and Fayette 1981, Hydrogeology of the Mam- county, C::entucky area: Lexing- moth Cave Region, Kentucky: Geo- THE INNER BLUE GRASS KARST REGION 6 1 1

logical Society of America Field the Inner Blue Grass Karst Trip ~uidebook, p. 457-506. Region, Kentucky: Field Guide for the Annual Meeting of the Quinlan, J. F., and Rowe, D. H.. Southeastern and North-Central 1977, Hydrogeology and water Sections, Geological Society of quality in the Central Kentucky America, 31 p. Karst: Phase I: University of Kentucky Water Resources Insti- Thrai 1ki l 1 , John, accepted for tute, Research Report 101, 93 p. publication, Flow in a limestone aquifer as determined from water Society of Dyers and Colouri sts, tracing and water levels in 1971, Colour index, (3rd ed.): we1 1s: Journal of Hydrology. 641 1 p. Thrai 1ki 11, John, Byrd, P. E. , Spangler, L. E., 1982, Karst hy- Hopper, Jr., W. H., McCann, M. drogeology of northern Fayette R., Spangler, L. E., Troester, and southern Scott counties, J. W. , Gouzie, D. R. , and Pogue. Kentucky: Lexington, University K. R., 1981, The Inner Biue of Kentucky, M. S. Thesis, 103 p. Grass Karst Region, Kentucky: An overview: Eighth International Spangler, L. E., Byrd, P. E., and Congress of Speleology Proceed- Thrailkill, John, submitted, Use ings, p. 336-338. of optical brightener and direct yellow dyes for water tracing in Thrailkill, John, and Gouzie, D. the Inner Blue Grass Karst R., 1984, Discharge and travel Region, Kentucky, Jones, W. time determinations in the Royal K. , ed. , Symposi um on water Spring groundwater basin, Ken- tracing techniques: National tucky: University of Kentucky Spel eological Society. Water Resources Research Insti- tute, Research Report 149, 43 p. Spangler , L. E. , and Thrai 1ki 11, John, 1981, Hydrogeol ogy of Thrailkill, John, Hopper, W. H., northern Fayette County and McCann, M. R., and Troester, J., southern Scott County, Kentucky, 1980, Problems associated with USA: Eighth International Con- urbanization in the Inner Flue gress of Speleology Proceedings, Grass C::arst Region (Abs. : Asso- p. 535-555. ciati on of Ameri can Geographers, Annual Meeting Program Ab- Thrai 1ki 11 , John, 1968, Chemical stracts, p. 112. and hydrologic factors in the excavation of 1imestone caves: Thrailkill, John, and Robl, T. L., Geological Society of America 1981. Carbonate geochemistry of Bulletin, v. 79, p. 19-46. vadose water recharging lime- stone aquifers: Journal of Hy- Thrai 1ki 11 , John, 1980, Inner Blue drology, v. 54, p. 195-208. Grass Karst Hegion (abs.): Pro- ceedings of the Kentucky Water Resources Institute, p. 25. Thrai 1ki 1 l , John, Spangler, L. E. , Hopper, W. M., Jr., McCann, M. Thrailkill, Johnp 1983, The nature R., Troester, J. W., and Gruiie, of groundwater basins in the D. R., 1983, Groundwater in the Inner Blue Grass Karst Region, Inner Blue Grass Karst Hegion, Kentucky (Abs. ) : Geological Soc- Kentucky: University of Kentucky iety of America, Abstracts with Water Research Institute, Re- Programs, v. 15, p. 97. search Report 136, 136 p. Thrai 1ki 11, John, Eyrd, S. B., Thrai 1ki 11, John, 1984, Hydrogeol- Sullivan, L. E., Spangler, L. ogy and environmental geology of E. , Taylor, C. J., Nelson, G. REFERENCES CITED

N. , and Prague, K. R. , 1983, geology: International Assacia- Studies in dye-tracing tech- tion of Hydrogeol ogi sts. niques and karst hydrogeology: University of Kentucky Water Re- U. S. Geological Survey, 1983, sources Research Institute, Water resources data for Ken- Research Report 140, 97 p. tucky: Water-Data Report KY- 83-1, 519 p.

Van Couvering, J. fl., 1962, Char- Thrailkill, John, and Troester, J. acteristics of large springs in W., 1978, Preliminary, supple- Kentucky: Kentucky Geological mental, and final reports on hy- Survey, ser. 10, Information drogeology of Georgetown, Ken- Circular 8, 37 p. tucky, in A report to assist

coal gasification demonstration I White, W. P., 1977, Role of solu- projection, Georgetown, Ken- tion kinetics in the develop- tucky: Lesington, G. Reynolds ment of karst aquifers, in Watkins, p. El-E14. Tolson, J. S., and Doyle, F. L., eds. , Earst hydrogeology: Twelfth Congress of the Inter- Thrailkill, John, Troester, J. W., national Association of Hydro- Spangler, L. S., and Cordiviola, geologists Memoirs, p. 503-517. S. J. , accepted for publication, Nature of a groundwater basin Whitesides, D. V., 1971, Yields divide near Georgetown, Inner and specific capacities of bed- Blue Grass Karst Region, Ken- rock wells in Eentucky: Kentucky tucky, USA, LaMoreaux, P. and Geological Survey, ser. 10, In- Burger, A., eds., Karst Hydro- formation Circular 21, 18 p. Chapter 4 PATTERNS OF CAVERN DEVELOPMENT ALONG THE CUMBERLAND ESCARPMENT IN SOUTHEASTERN KENTUCKY Ralph 0.Ewers Department of Geology Eastern Kentucky University Richmond, Kentucky

The Cave Creek area contains stones, sandstones, and shales of six known units of subsurface con- Upper Missi ssi ppi an and Lower duits totaling 17 km in length. Pennsylvanian age. Figure 1 shows These units form part of a once a representative stratigraphic fully integrated system of lime- section from the area. The nomen- stone caves. The inputs to this clature used here is that adopted system are confined to a narrow for the Shopville Quadrangle zone, no more than a few hundred (Lewis, 1972). meters across, where stream ero- Structure sion has breached an impermeable caprock consisting of sandstones The area lies on the eastern and shales. This system of con- flank of the Cincinnati Arch, a duits and inputs is relatively broad, low, north-south trending isolated from others and provides anticlinal feature that can be one of the clearest examples of traced from southern Michigan the most common pattern of cavern through Cincinnati to northern development in the region. Like- Alabama. On a regional basis, the wise, it provides an almost ideal rocks may, therefore, be described location to study the process by as dipping gently toward the east- which the caves of the region southeast at 3.5 to ? m per km. evol ve. Local 1y , however, a considerable amount of subtle structure can be GEOGRAPHICAL SETTING discerned. In the region of spec- ific interest, a sma.11 anticli- Cave Creek, a tributary of the nical feature extends over the point of conf 1uence between the Cumberland River, is located in Cumberland River and Cave Creek south-central Kentucky, in (Fig. 2). Pulaski County. The area is depic- ted on the Hail 7.5 minute geolog- Geomorphology ic quadrangle map (Smith and The region lies at the edge of others, 1'7731. The area is part of the mature1y d'issected western the larger karst region that margin of the Cumberland Plateau. extends from just south of the The plateau is capped with basal Ohio River, southward along the Pennsylvanian clastics of a high1y Appalachian Highlands Escarpment variable nature. They range from through Tennessee. cross-bedded conglomerates, local1 y referred to as the GEOLOGICAL SETTING Rockcastle Member by Hatch (1964), to siltsones, shales, and coals. Lithology The hilltops in the plateau region The region is underlain by sed- have an elevation of 375 m: imentary rocks, largely lime- maximum relief here is 155 m. HYDROLOGY

V) Kidder and Ste. Genevieve NOMENCLATURE 2 FORMATION V) V) limestones and is marked by C Y" OF EARLIER V) & LITHOLOGY >V) 5 5 numerous sinks and solution WORKERS MEMBER r I C valleys. To the west of the knobs . . -. . - . - .. - .. - is a gently rolling surface, . - .. - .. Z -. . -. . - 5 .. -. - -. continuous with that between the Z u knobs, which forms the Highland Rim, a part of the Interior Low Lee > V) Plateau Province (Fenneman, 1938). Z Z Y The Cumber 1and River f 1ows in a L narrow gorge for 64 km through the dissected portion of the escarpment. Evidence suggests that the gorge here was formed, in part, by the retreat of Cumberland Pennington Falls from near Burnside, Kentucky, to its present location upstream from the town. The falls retreats as it undermines the Glen Dean resistant Rockcastle Conglomerate

L caprock, exposing the easily 2 Hardinsburg Lo, eroded shales, siltstones, and Haney poorly cemented sandstones of the 0 Lee and Pennington formations Reelsville (McFarl an, 1943). The development Beech Creek of the gorge is closely related to the evolution of the karst in Cave Creek, as will be demonstrated be1ow. Paoli Beaver Bend The erosion of Cave Creek Valley has breached the caprack and exposed the Monteagle and Ste. Genevieve 1itnestones a1ong the valley sides and bottom. This Ste. Genev~eve exposure extends for 5.75 km eastward and southward from its ' a0 conf 1uence with the Cumber I and E m River L St. . Louis Cave Creek valley can be characterized as a karst valley, due to its irregular prof i 1e and its internal drainage. There are many other val 1eys of simi 1ar After Mcfarlan & Walker After Lewis. (1972) character along the escarpment. (1956) Hydrology Figure 1. Stratigraphy in the The drainage in the region is Cave Creek area, Kentucky. control led by the Cumberland River and, since 1752, by the artificial pool created by Wolf An area of knobs occurs in a Creek Dam. The normal pool narrow strip normally 5 to 6 km elevation for Lake Cumberland is wide along the edge of the 220 m, 5 m above the plateau. These detached plateau pre-impoundment level at the mouth remnants have summit heights up to of Cave Creek. The areas of about 365 m and rise about 100 m discontinuous sandstone and above the intervening surface. congl omerate act as granul ar This surface is developed on the aquifers, perched upon the

THE CAVE impermeable shales of the Lower from the region of the sink at "A" Pennsylvanian and Upper could not be conducted along path Missi ssi ppi an. These aqui f ers give "A-C" or on the direct course to rise to numerous diffuse seeps and "R. " No structural or 1i thologic springs. Where exposed, the barriers to such flow are known or limestone units develop subsurface seem likely. In fact, a conduit drainage. This secondary considerable advantage would seem solution porosity is not to exist along these latter restricted to any particular courses. Their lengths are 3 km stratigraphic horizon within the and 3.6 km, compared to 4.8 km. limestone, although there is some Thus, they should have 62.5 and 75 tendency for surface streams to be percent, respectively, of the maintained where the upper St. resistance of the longer course. Louis 1imestones 1ie direct1y The gradient would be 1: 3 and beneath the surface. This tendency 1:3.6, respectively, compared to suggests that this unit may be 1:4.8, a significant increase. less soluble, or otherwise less Stratigraphical1 y, point "C" is conducive to karstification. about 2 m lower than point "P," Cave Creek Val ley contains and, therefore, if the several hundred closed depressions entrenchment of the Cumberland and is without continuous surface River proceeded uniformly, point flow, even under conditions of "C" may not have been able to extreme precipitation. fit least 30 discharge waters from the well defined wet-weather 1imestone as early. tributaries, carrying allogenic In the face of these waters from the caprock, extend to theoretical advantages, the known the valley bottom, where they sink conduits follow the longer, lower (Fig. 2). The resurgence for these gradient course. It seems waters is assumed to lie at the reasonable, given the geomorphic Cumberland River beneath its setting, that the f 01 lowing events present artif icia1 pool. A large would have occurred. As the falls spring, downstream from the valley receded past the mouth of Cave mouth and on the same side of the Creek Valley, the stream flowing river, appears on early on the caprock sequence began to topographic maps (Mayfield and gradually expose the limestone Withers, 1929) and is reported by along Cave Creek to significant long-term residents of the area. hydraulic pressure gradients. This This spring is presumed to be a exposure, by reason of the dip as r.esurgence. well as the gradient of the stream, would have proceeded from the valley mouth in an upstream The Cave direction. The localization of the The known passages associated first breaching at the creek with Cave Creek Valley form a mouth, rather than at other points three dimensional network that along this reach of the river, is closely f 01 lows the valley axis assured by the presence of a minor (Fig. 3a). The conduits are formed anticline at this point (Fig. 2). along bedding planes and are Second, upon the exposure of a virtual1y without joint control. suitable structural discontinuity The network ranges vertically for admitting water to the between 207 m and 263 m. The limestone bedding planes, the course of the subsurface flow development of one or more whi ch traversed these c-onduits, subsurf ace condui ts between this neglecting minor deviations, input and the Cumberland River extends for a distance of 4.8 km. would have occurred. Third, as Figure 3b depicts this generalized additonal inputs evolved, they path as solid line "A-R." There is developed conduits linking to the no reason to believe that flow previous1y completed conduits. PATTERNS OF CAVERN DEVELOPMENT ALONG THE CUMBERLAND ESCARPMENT 67

Kilometres

Figure 3-A, E. Conduit pattern and generalized subsurface flow route beneath the Cave Creek area. Cave inforrnation was provided courtesy of Mr. Louis Simpson. THE CAVE

Kilometres

Figure 3-A, B. Conduit pattern and generalized subsurface flow route beneath the Cave Creek area. Cave information was provided courtesy of Mr. Louis Simpson. PATTERNS OF CAVERN DEVELOPMENT ALONG THE CUMBERLAND ESCARPMENT 69

In support of this scenario are on1 y af ter a conduit had three pieces of evidence. Fir~t , established a low-resistance 1ink only 4 km of passage, less than 25 to its discharge target. The percent of the total known, sediments armor the lower portion extends beneath the caprock on the of the conduit and concentrate valley flanks. This scarcity dissolution at the ceiling. indicates that the targets for Frequently, the passage shaws conduit development 1ay a1 ong the meander forms that propagate valley bottom. Second, there are upward and downstream, simi lar to several regions of conduit the presumed growth of eskers in directly connected with the trunk glacial ice (Embleton and King, conduits that presently lie close 1971, p. 370-382), thus verifying to the surface and near the valley the presence of this mode of center. These regions appear to en1argement . The donstream have functioned as specific sites direction of flow is deduced from of input. The most complicated scallops on the walls, the mazes are located in the upper imbrication of gravels where they part of the val 1ey, suggesting exist, and from the assumption that those in the lower part, and that flow was from the valley so formed earlier, may have been inputs toward the major trunks. partly destroyed. Third, many of fill of these methods give a the high elevation mazes are in consistent flow direction. This the form of phreatic canyons. The mechanism has been discussed by morph~logyof these conduits Ewers (1977) , Passini ( 1973) , who suggests that they have enlarged used the term "antigravitational upward from a horizontal bedding erosion, " and Renaul t ( 1967) who plane network (Fig. 4) or from a used the term "paragenesis." This descending primitive tube network, latter evidence strongly suggests following a combinaton of joints that the main conduit development and bedding planes. This was truly phreatic, not simp1 y enlargement normally occurs in a epiphreatic. Furthermore, it turbulent regime when sediments supports the contention that the are carried into the conduit. The mazes were input points. enlargement would have occurred The passages illustrated in Figure 3 represent two fairly distinct levels of conduit development, one betweeen 200 m and 225 m, and another between 225 m and 240 m. Although these ranges overlap, the juxtaposition of passages confirms that they are parts of two distinct trunks. This distinction suggests that development of primitive phreatic tubes may have occurred at several horizons, with the upper level enlarging to trunk proportions first. Later, when the river had entrenched further, the lower level enlarged and became active.

A THEORETICAL CONSIDERATION Figure 4. Phreatic Canyon from Goldsons Cave, Cave Creek, Ken- The scenario presented above is tucky. This conduit is inferred consistent with anal yses of cavern to have formed upward from the evolution based upon three bedding-plane network under fundamental factors: the porosity phreatic conditions. distribution, the head THE SOLUTION KINETICS

distribution, and the solution the region 3a - 3b boundary. He kinetics. The porosity then calculated the travel distribution sets strict limits on distance required to reach 90 the location of the initial percent saturation for a range of solution porosity. The head capillary openings and groundwater distribution within this porosity flow gradients (Fig. 6). For this, controls the rate and direction of he used the rate equations of the initial groundwater movement. Plummer and Wigley (1976) which The rate at which the moving suggest a second order surf ace groundwater approaches saturation reaction control, and the may have a profound effect upon diffusion control model advocated the pressure distribution as the by Weyl (1958). He selected carbon solution porosity evolves. If the dioxide partial pressure values of kinetics are fast, undersaturated 10E-2.5 and 10E-1 for the water will emerge from the reaction-1 imi ted calculations. limestone mass and the fracture These values correspond to typical porosity, which conducts this conduit spring values and typical flow, will enlarge uniformly soil-water values for carbon throughout its length. On the dioxide concentrations, other hand, if the kinetics are respectively. These plots cover slow, saturated water will be the full range of pertinent, discharged and the fracture pub1 ished, experimental data and porosity will enlarge initially in theoretical arguments. As White the region of the input and will pointed out, he substantially qradually propogate toward the agrees with Perner and Morse that discharge region. Such a for small capillaries and non-uniform change in the porosity fractures representative of virgin will alter the pressure pore space of tectonic, distribution in the surroundr ng diagenetic, and sedimentary origin limestone, thereby affecting in limestone, the bulk of the further change. geomorphic work accomplished by the dissolution process takes 1 Solution Kinetics place within a few meters or tens Figure 5 shows the dissolution of meters of the point where rate experiments of Berner and groundwater first gains access to Morse (1974) plotted as functions the limestone. of the deviation of pH from its In a more recent series of equilibrium value. The plot can be calcite dissolution experiments, readily divided into three regions Plummer and others ( 1978) have covering five orders of magnitude. extended the pH, carbon dioxide Each of these regions has a partial pressure, and temperature characteristic rate of change in range of previous experiments. the solution rate. Of particular They proposed a mechanistic model interest here is the dramatic drop that describes their observations. in the dissolution rate in region Their results, in general, 3. . Thi s drop suggests that corroborate the results of Berner aggressive (undersaturated) water and Morse ( 1974) . might penetrate to great These laboratory investigations distances, but the rate at which are supported by three types of it accomplishes geomorphic work field observations. First, Bog1 i should be quite small except near ( 1966) and Cog1 ey ( 1972) have the point of input. demonstrated that thin films of White (1977) reviewed the water from rain and snowmelt Berner and Morse results in a traversing exposed limestone karst context. He pointed out that surfaces quickly saturate to for karst waters the break in levels appropriate for open-system slope occurs at about 90 percent waters in contact with atmospheric saturation, which corresponds to concentrations of carbon dioxide. PATTERNS OF CAVERN DEVELOPMENT ALONG THE CUMBERLAND ESCARPMENT . 71

_-----

I I I Log C02 Pressure = 0.0 I I Log C02 Pressure = -1.5 I A Log C02 Pressure = -2.0 I 0 Log C02 Pressure = -2.5 I , Log C02 Pressure = -26 I A'Log C02 Pressure = -3.5 I

I

I I I I

I

Figure 5. Solution rate experiment data of Berner and Morse (1974) plotted as functions of the deviation of pH from its equilibrium value at the given carbon dioxide partial pressure (from White, 1977).

Second, in the subsurface, large cavities are only a few meters numbers of "" beneath the surface. These form from seeps along deposits form by crystal 1 ization cavern ceilings. Frequently, these at the perimeter of successive 7 2 THE POROSITY DISTRIBUTION

yet they have achieved a degree of saturation that not only makes growth possible but precludes the re-solution of its base from the inside in these slender forms (Ewers, 1982) . Finally, White (1977) pointed out that springs in the Appalachians are typically undersaturated at the critical 90 percent level. This undersaturation corresponds to the dramatic decline in the mass transfer rate in region 3 of the Ferner and Morse experiments.

Porosity Distribution

Secondary permeability, in the I form of joints and bedding planes, should provide the capillary spaces through which the bulk of groundwater movement in limestone will occur. In support of this statement is the widely published data, from many sources, on the low primary permeability of I I I I I Paleozoic 1i mestones with which 0.01 0.1 I 10 1 O(3 this study deals. For example, CAPILLARY RADIUS (Cm) Davis and Dewiest (1966, p. 348), Choquett and Prey (1970), and Figure 6. Distance to 90 per- Freeze and Cherry (1979, p. 29) cent saturation with respect to listed limestone permeabilities of calcite as a function of conduit less than 0.1 millidarcies when diameter. Hydraulic potential fracture porosity is not surface slopes of 1:50 and considered. Early workers such as 1:5,000 are shown, calculated Martel 1 (1921) and Swinnerton with the Plummer and Wigley (1932) pointed out the importance (1978) rate equations. A slope of these partings. 1:500 is also calculated using Bedding planes have been shown the Weyl (1958) model. Carbon by several authors to be dioxide partial pressures of associated with phreatic conduits, 10E-2.52 and 10E-1 are used, re- and quantitatively more closely flecting values typical of associated with these conduits spring waters and soil waters, than joints. Ewers (19721, in an respectively. (Adapted from analysis of more than 20 km of White, 1977.) subsurf ace conduits in south-central Kentucky, showed that 93 percent of these were apparently related to bedding planes. Deike (1967) found joints drops of seepage water supplied of very limited importance in the through a central canal. devel opment of subsurf ace karst in Typically, soda straws are a few the Mammoth Cave Region, and by decimeters in length but range to implication, that bedding planes 1 m or more. The shallowness of are of great importance. Ford some of these caverns makes it (1971) reported that, in his field probable that the drip waters have experience, the ratio of bedding not travelled far Srom their plane to joint passages commonly initial contact with limestone, ranges from 10: 1 to 100:l in those PATTERNS OF CAVERN DEVELOPMENT ALONG THE CUMBERLAND ESCARPMENT 73 caves where bedding planes are of bands (Ford, 1968). any importance at all. 3. The rate and direction of that The reasons for the propagation is related to the overwhelming importance of bedding pressure field within the bed- planes in this regard is not ding plane. This pressure is, difficult to understand. Bedding in turn, related to the geome- planes are frequently more common try of the inputs and resur- than major joints, and their gences and their relative iluid lateral ex tent is of ten much potentials. greater, often reaching all 4. As the networks grow, the pres- boundari es of a 1i mestone mass. sure field must change, and Joints that do traverse an entire networks possessing high growth rock mass can carry rate will retard the growth of on1 y along a single horizontal those with slower growth rates. vector, which may not coincide 5. The hydraulic capacity of a with the regional hydraulic tube network changes abruptly potential field. Bedding planes when it fully penetrates a are capable of conducting flow beddinq plane. along any horizontal vector. Even Prior to the establishment of a in those cases where complementary low-resistance 1ink with the joint sets may provide a output, the meteoric water continuous capillary opening to catchment for an input is likely the limestone boundary, it is not to be small, because the network's clear that it will be enlarged. discharge capacity is small. The Ford (1971) pointed out that discharge capaci t y is determined groundwater flowing through a by the transmissivity of the network of joints is required to remaining una.1tered bedding plane make many turns at joint through which the growing tube intersections. Such a course is network must ,discharge, not the one of high resistance and network itself. When a network vulnerable to capture by a breaches the bedding plane, it may straighter, more efficient bedding then be capable of conducting a1 1 plane, with which the joint almost of the water available to it. In certainly intersects. such a case, the head throughout the network.wi 11 approach the A THEORETICAL MODEL level of the resurgence. This will produce a depression in the Arguing from these stated hydraulic potential field in the principles and the generally region of the network. Surrounding accepted principles of flow in networks will respond to this porous media, five statements pressure field change with an logically follow. increase in their growth rates and 1. Because the solution kinetics a redirection of their growth are quite rapid, the initial toward the 1ow-resi stance tube. solution porosi t y wi1 1 propa- The first of these to link with gate from the input toward the the initial low-resistance tube resurgence. will become the dischrge target 2. Because the pore space conduct- for some of the remaining ing the groundwater flow, a networks. These remaining netwoks, bedding plane, is a thin three- in turn, become the targets for dimensional space, each input additional networks developing will discharge along a separate from still more distant inputs. vector. Thus, each vector will Thus, the cave grows in length by give rise to a separate network the linking of short segments of of tubes. Ewers (1982) has condui t in a headward direction shown that these networks and a stepwise fashion. The gross should be in the form of dis- pattern is tributary in form, but tributaries or anastomotic the individual elements are EVOLUTION OF THE LINKING PATTERN distributaries. This evolutionary the resurgence, and the network scheme has been called the associated with it will function progressive headloss model (White, as a sink for groundwaters 1977) . The detai 1s of any entrained in the bedding plane. particular linking system depend The discharge boundary is thereby upon boundary conditions imposed altered from a point with limited by topography and geologic capacity to one of linear structure. Three linking models dimension, equal to the perimeter have been investigated: high-dip, of network 1, with greatly low-dip, and restricted-input increased capacity. In response to (Ewers, 1982). The latter applies this change and its newly most commonly to the Cumberland established proximity to an output Escarpment karst and to Cave Creek boundary, network 2 should specifically. commence rapid growth. The new growth of this network will take Evolution of the Linking Pattern place along a route that combines already existing tubes with the Figures 7a-e i1 lustrate the shortest available path to form a development of the restricted link with network 1 (Fig. 7d). input linking pattern. In this Network 3 should show increased idealized case, three inputs and a growth following the breaching of single-point resurgence occur along a stream course where an the bedding plane by network 1. impermeable caprock has been Subsequent to the linking of breached. Groundwater is assumed networks 1 and 2, network 3 should to be conducted to and from the commence growth to complete a low bedding plane by way of joints. resistance link with network 2, in Equal driving potentials hl, the manner already described (Fig. 7e) h~,and h= are assumed . for the inputs. Porous media Laboratory solution experiments hydraulics require that input 1 verify that there is good reason will have the highest throughput, for such a linking system to be with successively small er Q values opera.ble in nature: it is simply for Inputs 2 and 3. The solution more efficient. Where the kinetics predict that the growth possibility exists to drive a rates of the tube networks arising groundwater conduit over a given from the three inputs will be in distance from a single input or proportion to their throughputs. from several intermediately spaced The flow from input 1, the inputs in a linking scheme, the proximal input, is symmetrical multiple input systems will prove about the input-output axis, and more time-efficient (Ewers. 1982). the flow envelope will possess an ovoid shape. A tube network should develop along the center of the SUMMARY initial flow envelope. Inputs 2 and 3, however, exhibit a The restricted-input linking two-1 obed f 1ow envelope and scheme can be characterized in the initially generate a pair of tube f 01 1owing ways: networks that form relatively 1. Subsurface drainage basins independent1y (Fig. 7b). should be established by the In Figure 7c the network stepwise integration of small forming from input 1 has completed distributary sub-units into a a low-resistance connection to the tributary system. output, and it is assumed to have 2. Similarly, the integration of enlarged to a diameter sufficient the system should proceed from to conduct all of the water the resurgence in a headward available at Input 1 under average direction, the opposite of the conditions. Therefore, the head at direction of the network input 1 will drop to the level of propagation. PATTERNS OF CAVERN DEVELOPMENT ALONG THE CUMBERLAND ESCARPMENT 75

Figure 7. Stages in the development of the restricted-input linking pattern. 7 6 REFERENCES CITED

3. The plan form of the conduit Dei ke, G. H. , 1967, The develop- pattern should be simple, lin- ment of caverns of the Mammoth ear, and relative1y unbranched Cave Region: University Park, and should parallel the stream Pennsylvania, Pennsylvania State course from which it University, Ph. D. Dissertation, originated. 235 p. The morphology of the conduits beneath Cave Creek Valley are Embleton, C. , and King, C. A. M. , consistent with their having 1971, Glacial and periglacial evolved in the manner described in geomorphology: Toronto, Mac- the restricted-input model. The Mi1 lan of Canada, 608 p. fact that alternate, but unused, pathways of steeper gradient for Ewers, R. O., 1972, A model for discharge of the headwaters of the the devel opment of subsurf ace system exist argues strongly that drainage routes along bedding the theoretical growth advantages planes: Cincinnati, Ohio, Univ- of the multiple input system are ersity of Cincinnati, M.S. real. This mode of cavern Thesis, 84 p. development is the most common type in the region of the Ewers, R. O., 1977, A model for Cumber 1and Escarpment. the development of broad scale networks of groundwater flow in carbonate aquifers, jn Tolson, J. S. , and Doyle, F. L. , eds. , eds. , Karst hydrogeology: Hunts- ville, Alabama, University of A1 abama at Huntsvi 11e Press, REFERENCES Memoirs, v. XII, p. 503-517.

Eerner, R. A., and Morse, J. W., Ewers, R. O., 1982, Cavern devel- 1974, Dissolution kinetics of opment in the dimensions of in sea water: length and breadth: Hami 1ton, IV. Theory of calcite dissolu- Ontario, HcMaster University, tion: American Journal of Sci- Ph.D. Dissertation, 398 p. ence, v. 274, p. 108-134. Fenneman, N. M., 1938, Physiogra- Eogli, A., 1966, Karstwasserflache phy of the eastern United und unterirdische Karstniveaus: States: New York, McGraw-Hill Erdkunde, v. 20, p. 11-19. Book Co., 714 p.

Choquett, P. W., and Prey, L. C., Ford, D. C.. 1968, Features of 1970, Geologic nomenclature and cavern development in central classification of porosity in sedimentary carbonates: American Mendip: Transactions of the Cave Associ ati on of Petr01 eum Geol o- Research Group of Great Britain, gists Bulletin, v. 54, no. 2, p. v. 1.0, p. 11-25. 207-250. Ford, D. C., 1971, Geologic struc- Cogley, J. G., 1972, Processes of ture and a new explanation of solution in the Arctic limestone 1imestone cavern genesis: Trans- terrain, in Polar Geomorphology: actions of Cave Research Group Institute of British Geography, of Great Britain, v. 13, p. Special Publication, No. 4. 81-94.

Davis, S. N. , and Dewiest, R. J. Freeze, R. A., and Cherry. J. A., M. , 1966, Hydrogeology: New 1979. Groundwater: Engl ewood York, John Wiley and Sons, Cliffs, New Jersey, Prentice- 463 p. Hall, Inc., 604 p. / PATTERNS OF CAVERN DEVELOPMENT ALONG THE CUMBERLAND ESCARPMENT 77

Hatch, N. L., Jr. , 1964, Geology Plummer, L. N. , Wigley, T. M. , and of the Shopvi 11e Quadrangle, Parkhurst, D. L., 1978, The ki- Kentucky: U. S. Geological Sur- netics of calcite dissolution in vey Geology Buadrangle Map COa-water systems at 5" C to 60- CQ-282. C and 0.0 to 1.0 atm. CO,: American Journal of Science, Lewis, R. Q., Sr. , 1971, The Mont- 278, p. 179-216. eagl e Limestone of south-central Kentucky: U. S. Geological Sur- vey Bulletin 1324.E. 10 p. Renault, P., 1967, Le probleme de 1a spel eogenese: Anna1 es de Martell, E. A., 1921, Nouveau Speleologie, v. 22, p. 5-21, traite des eaux souterraines: 209-267. , , Editions Doin, 840 p. Smith, J. H., Pomerene, J. B., Ping, H. G., 1973, Geologic map Mayf ield, S. M. , and Withers, S. , of the Hail Quadrangle, McCreary 1929, Map of areal and structur- and Pulaski counties, Kentucky: al geology of Pulaski County, U. S. Geological Survey Geologic Kentucky: Kentucky Geological Quadrangle Hap GG!-1058. Survey, ser. 6, scale 1:62,500.

McFarlan, A. C., 1943, Geology of Swinnerton, A. C., 1932, Origin of Kentucky: Lexington, University 1i mestone caverns: Geol ogical of Kentucky, 531 pp. Society of America Bulletin, v. 43, p . 662-693. Passini , G. , 1973, Sull 'importanza spel eogenet ica dell ' "erosi one Weyl, P. K., 1958, The solution antigravitativa": Estratto da Le kinetics of calcite: Journal of Grotte D'Italia, Serie 4, v. IV, Geology, v. 66, p. 163-176. p. 297-322.

Plummer, L. N., and Wigley, T. M. White, W. E., 1977, Hole of solu- L., 1976, The dissolution of tion kinetics in the development calcite in COz-saturated of karst aquifers, in Tolson, J. solutions at 25" C and 1 S., and Doyle, F. L., eds., atmosphere total pressure. Karst Hydrogeology: Huntsville, Geochim. Cosmochim. Acta, v. 40, Alabama, University of Alabama p. 191-202. at Huntsville Press, p. 503-517. Chapter 5 CAVES OF NORTHEASTERN KENTUCKY (WITH SPECIAL EMPHASIS ON CARTER CAVES STATE PARK) John Tierney Park Naturalist Carter Caves State Resort Park Olive Hill. Kentucky 41164

East on Interstate 64 from the 1if t occurred, erosianal forces heart of the Blue Grass area, the kept pace, wearing away the arch jagged rock outcrops are passed so as it pushed upward. Today, there quickly, it is difficult to appre- is no mountain grandeur in central ciate the incomprehensible amount Kentucky--just the core of the of time that led to their forma- arch exposed. In the center of the tion. It takes only about 2 1/2 uplift, the youngest rocks have hours to travel the 193 km from been eroded away, leaving the Lexington to Ashland, in the older rocks exposed. As one northeastern corner of the State. travels from the center of the Put in that stretch, one passes Blue Grass, progressive1y younger rocks that took over 400 million rocks outcrop like concentric years to accumulate. growth rings in the center of a The Paleozoic Era, 600 to 250 tree trunk. million years ago, was a time when A caver, traveling east on all of Kentucky was under water. Interstate 64 from Lexington, Sediments accumulated on the wauld not have much to be floors of these ancient seas and enthused about for some distance. eventually were compressed to form One would first pass the thin- the rocks we now see. We know it bedded, fossiliferous limestones was a time of regression and of the Ordovician Period. Not much transgression of these ancient in the way of caves there. Farther seas. There were times of tran- along, one would pass the shales qui 1i ty , when ani ma1 remains and dolomites of the Silurian and could easi 1y be preserved. There Devonian periods. Nothing there. werg times of turbulence, when the In this area, a chain of conical- remains were ground up into micro- shaped hills can be seen. These 'scopic bits and pieces. All of are the Knobs. They form a ring these paleoenvironmental fluctu- around the Blue Grass. The ations are readily observed due to Missi ssippi an and Pennsyl vani an the diversity of rock type stacked rocks are exposed at the tops of in layers, one on top of the these Knobs, and it is within the other, from the oldest to the next 40 km of Interstate highway youngest. that one would cross a northeast- At some point in time after the southwest trending be1t of 1ime- rocks had been formed, there was a stones in ,which most of the caves period when these rocks were of northeastern Kentucky may be thrust vertically upward along a found. line running north-south from The Mississippi an 1i mestones northwestern Ohio through exposed in this region have been Cincinnati to the Nashville, identified as the same limestones Tennessee area. Had subsequent found in similar stratigraphic erosional events not af f ected thi5 position in other parts of the Cincinnati Arch, there would be a State. However, in northeastern high ridge in central Kentucky Kentucky, the Mississippi an and today, instead of rolling, pastor- Pennsylvanian periods were times al farms and fields. As this up- of transgression and regression of CAVES OF NORTHEASTERN KENTUCKY

inland seas. Water inundated areas attributed to glacial events that for a time and then receded. reduced the erosional energy of Barrier isl ands of sand developed, the streams in the area. Tygart's only to be covered by Creek is a fine example of a transgressing seas. Limestones stream that has been rejuvenated. interface with sandstones to show Once it meandered over a wide the transitional energies that area, slow1y cutting through the were occurring at the time. In the rock. Then with glacial melting region, the maximum thickness of and a more easily eroded good homogeneous limestone substrate, the stream began a available for solution is 27 m. periad of entrenchment which is Tygart's Creek is a relatively clearly visible in the vicinity of small stream that meanders Carter Caves State Resort Park. northeast through Carter and Today along the northern Greenup counties on its way to the Cumberland Plateau, the tops of Ohio River. Through a 32 km the hills are usually around 305 stretch, it crosses the limestones to 427 m in elevation, while the of northeastern Kentucky, and valley floors are usually 152 to within that section most of the 213 m in elevation. The signi f icant cave devel opment of cave-forming rock, the Newman the region occurs. The geologic limestones, outcrop near the 244 m history of this stream and its contour level and extend downward. tributaries is vital1 y important The erosional progress of Tygart 's to the understanding of the Creek has taken it through the spel eogenesi s. Newman sequence to the thin, Prior to the glacial and impure 1imestones correlated to interglacial events of the the St. Louis Formation. Below Pleistocene, Tygart's Creek, then this, the shaly, silty strata of a part of what has been called the the Borden Formaton are not Teays River drainage system, was readily dissolved by surface or eroding downward toward the subsurface water. Thus, Tygart's Mississippian 1imestones that had Creek itself is no longer a part been deposited millions of years of the cave-forming process. earlier. As the stream eroded However, tributaries are still downward, the limestones became eroding through the Newman, and more vulnerable to the chemical this is where the subsurface water and mechanical actions of the routes still exist. Within the watershed. Water, seeping into area, many springs of water emerge cracks and crevices of the at the contact point of the Newman limestone, began the process of limestones and the less soluble solution along joints and bedding strata beneath it. planes. During the early glacial Within the northeastern region events of the Pleistocene, many of of Kentucky, Carter Caves State the cave passages found in the Resort Park is the best known and upper portions of the limestones most intensive1y visited cave were being acted upon. With site. The Park is located about 8 subsequent interglacial periods, km off Interstate 64 in Carter Tygart's Creek and its tributaries County, Kentucky (the origin of experienced periods of rapid down the Park's name). The relief cutting. By the end of the last within the Park averages 152 m. At glacial transgression, most of the the highest elevations within the caves in the area had achieved Park, the orthoquartzitic much of their present size. During sandstones predominate the rock the last glacial event, many of outcrops. The tidal barrier the passages in the caves were deposits intergrade in other parts practically filled with outwash of the region with silts and sandy and sediments. Terracing of the 1i mestones of the Pennington slopes in the region can be Formation. Of notable interest to CAVES OF NORTHEASTERN KENTUCKY the serious visitor is the parallel joints that were erosional arches and rock houses dissected by a third joint running found in the sandstones. The more essenti a1 1y perpendicular to the interesting of these are Fern others. The result is a feature Bridge and Haven Bridge. Fern with two nearly perfect 90 degree Bridge is an arch believed to have corners. been formed by a . The There are approximately 20 water at some point in time was caves that are collectively called diverted through a crevice in the the Carter Caves. NO one cave outcrops and subsequently undercut bears the name. They range from the upper rock strata to produce 3.35 km down to those that could the bridge as it appears today. barely be called true caves. Haven Bridge is a much smaller Currently, the largest is Bat arch with well defined openings on Cave (Fig. 1). This is an both sides of the arch. Water, undeveloped cave that is best undercutting the rock, and known as a hibernaculum for the subsequent col 1apse of additional endangered Indiana bat 1Myotis strata, has led to the formation sodalis). This is the third of this delicately balanced arch. largest colony of this bat in the Another feature of interest in world, and because of the the sandstones of Carter Caves is vulnerability of the Indiana bat a feature called Box Canyon. to wintertime disturbance, the Located in a remote section of the cave is gated. Tours are conducted Cascade Area, the surface feature through the cave by Park staff was produced by collapse along during the summer months when the

BAT CAVE UNDERGROUND PASSAGE OF CAVE BRANCH

WILLIAM P EIDSON, JR MAY lVbb

100 2?01ttl I > F'>Lp -

Figure 1. Bat Cave, the longest cave in Carter Caves State Resort Park and a hibernaculum of the Indiana bat, Myotis sodalis. CAVES OF NORTHEASTERN KENTUCKY 8 1

bats are not in the cave. Fat Cave present-day stream bed exists. is made up of two levels of Cascade Cave, formed by James passages running parallel to one Eranch during its subterranean another. The main, 1ower-1 eve1 journey, was commercially passage is a wide, underground developed in 1925. It is conduit that was formed by considered by most to be one of solution along bedding planes. the most scenic caves in the area. Large rooms along this passage, The four main passages run where the ceiling reaches 10.7 m parallel at slightly different above the floor have resulted from elevations. The size of the the collapse of the thin rock passages along-the tourist route strata that forms the ceiling. A makes this bedding small stream, Cave Eranch, flows plane-control led cave seem quite through this passageway, making it large, if not in length, certainly susceptible to flooding during in the volume of the cavity. At periods of heavy rainfa1 1. Debris, main intersections of passageways, gravel, and mud.-caked surf aces attest to this. The upper levels of the cave are drier, and some growth has occurred. Fat Cave was uncontrolled up to 1974. Bef ore that time, thousands of people roaming through the cave extensively damaged many of the formations. Today, much of the original beauty of the cave is gone because of the CASCADE thought1essness of earl ier cave \\ \ \ CAVERNS visitors. A project is currently underway, and Eat Cave is likely to be in excess of 3.05 km in total length. Another cave in Carter Caves State Resort Park is Cascade Cave (Fig. 2). A small stream named James Branch begins inauspiciously alongside Interstate 64 and travels in a norther1y direction toward Tygart 's Creek. Flowing through the hilly farm country, 3.2 km into its journey, it drops over a waterfall approximately DOME RooM~V~"""OF EDEN 12.2. m high. The resistant rock UNDERORWIND EXITMA that causes the falls is the area's sandstone which sits on top of the Newman limestones. As the water drops over the falls it enters a valley underlain by the limestone, and shortly, the water flow has been diverted into the cracks of the rock and the streambed is dry. Under this val 1ey f 1oor , ex tensi ve cave cdevelopment has occurred. Figure 2. Sketches showing the Throughout the remainder of its main passages in the two lighted course, James Branch remains below commercial caves at Carter Caves the surf ace. The valley above is State Park. (from Geology of pitted with sinkholes, and no Carter Caves State Park) 82 CAVES OF NORTHEASTERN KENTUCKY enlargement has occurred through plane, has been diverted downward collapse, and at one of these through a joint to another bedding points called the Lake Room, the plane water course. The visitor visitor can see James Branch exit can view the 11 m plunge from a the cave through a large opening viewing platform constructed at to the outside. The water pools in about the midpoint of the falls. the opening, and an impressive The water eventually reaches James mirror lake that reflects images Branch within Cascade Cave, but of the landscape outside the cave attempts to physically connect the is created. In a disjunct part of waterfalls to the main cave have, the tour, the Cathedral Room can so far, been unsuccessful . Within be seen. Underneath a massive the last few years, area cavers system of sinkholes, large amounts working in conjunction with Park of water seep into the cave and off ici a1 s have surveyed regions o+ deposit a number of formations. the cave not previously connected Many columns have developed in the to Cascade Cave. They are hopeful room. One of them, resembles a that this cave may become the or cathedral sitting on a longest in the Park. hill, and thus gives rise to the The most unusual, and perhaps name of that section of the cave the most interesting cave of (Fig. 3). Carter Caves is Sal tpetre Cave. Another disjunct part of the Formed along a bedding plane near commercial tour is the Underground the top of the Newman Limestone, Waterfalls. It is located about this cave's broad passages are 91.4 m west of a sinkhole entrance mostly dry and dusty. The dirt and to Cascade Cave. Access to the gravel that near1y fill the waterfalls is through a man-made passages in many places show that entrance. This interesting the cave has not always been the phenomenon is a dome-pit structure way it looks today. The black soot like those of many other areas. that covers the walls of the cave Water, flowing along a bedding and the old tools and devices found in the cave suggest that the cave has some historical significance (Fig. 4). No documentation exists concerning the activities that took place in

Figure 3. The Castle in Figure 4. Signatures of early Cathedral Room of Cascade Cave. visitors in Saltpeter Cave. CAVES OF NORTHEASTERN KENTUCKY this cave, but legends told in the tall and over 2.2 m in diameter is area suggest that the dirt on the called the Giant Column. A series floors of the cave was mined as of drapery type formations is early as the War of 1812 for the called the Pipe Organ because of extraction of sodium nitrate the early custom of tapping .the (saltpetre) . The extreme dry formation to produce musical tones conditions of the cave are due to (a custom which is no longer the impermeable nature of the demonstrated). sandstone that covers practicallv Two other caves worthy of note the whole cave. Near the main at Carter Caves are Laurel Cave entrance to the cave, where the and Horn Hol low Cave. Laurel Cave sandstone does not overlay the is an undeveloped cave that is cave, dome-pits have developed by formed along a vertical joint. The the invading water from the downstream entrance is located surface. In a few places of this alongside Cave Branch. The part of the cave, mineral-laden upstream end of the main passage waters have been able to deposit is in Horn Hollow, an elevated some calcite formations. What the valley that seldom has water visitor sees in Saltpetre Cave, flowing along the valley floor. more than anything else, is the Instead, the water i5 underground. dark, dingy color of the walls and Further upstream, Horn Hollow Cave formations due to the flames of may be seen. Water flows out of torches and lanterns used in the the entrance, forming a large pool cave in earlier days. Recent that necessitates a chilly wade to surveying of this cave has put the enter the cave. This is strictly length at just under 3.05 km. an underground water conduit for Lantern-lit tours are conducted Horn Hollow Creek. The volume of daily throughout the year. During water that occasionally flows down the summer, cave guides also this valley makes this a cave to conduct frequent spelunking trips avoid dur i ng per iods of heavy through the cave by advanced rainfa1 1. reservation. Within a 40 km radius of Carter Just across the parking lot Cave there may be as many as 200 from Saltpetre Cave is X Cave. named pits and caves. To the Electrically lighted, it offers an northeast of Carter Caves, another extreme contrast to Saltpetre cave system of note is what cavers Cave. X Cave is a cave of narraw, have called The Cow high-cei 1i nged passages. It has Counterfeiter's Cave System. two main passsages formed along Around the turn of the century, vertical joints, tht intersect in parts of this cave were developed the middle of the cave, so that for tours by private owners. They when standing at the intersection, called it Oliginook Caves. The there is an obvious "Xu cave was shown for a brief time, configuration of the passages. but since then it has become a Each trunk of t.he "X" has an favorite with local cavers because outlet. One side of the cave is of the intricate maze of crawlways under the sandstone caprock, which and climbs. This system has formed makes it drier and, for the most along a series of part, except where dome-pi t northeast-southwest trending activity exists, void of joints. There are three sizeable formations. The other side is wet entrances to the cave. The and has a great deal of on-going connecting passages from one joint speleothem growth. The more to the next are small, and significant formations have been crawling and some technical climbs named over the years. A massive are necessary to visit all parts col lection of columns, of the cave. All of the entrances stalactites, and that are located near the top of a collectively form a deposit 9.1 m formidable hill. The walk to the 84 REFERENCES CITED cave from the car weeds out the REFERENCES weaker cave visitors. Dever, G. R., 1980, Stratigraph- Because the limestone thickness ic relationships in the lower is less than in the more and middl e Newman Limestone significant cave areas of the (Mi ssissi ppi an) east-central State, there are few caves which , and northeastern Kentucky: Ken- attract the vertical caver to tucky Geological Survey, 49 p. northeastern Kentucky. There are a number of pits in the area, but Dever, G. R., Hoge, H. P., Hester, the deepest of them would not be N. C., and Ettensohn, F. R., much deeper than 24.4 m. 1977, Stratigraphic evidence for Because much of the solution late Paleozoic tectonism in and development of caves has Northeastern Kentucky, iq Field occurred along joints, many of the Trip Guide, for the Eastern caves offer some complexity to Section, American Association of rock climbing and negotiating Petroleum Geologists, 80 p. through narrow, confining passages. One cave that offers a Faiio, V. and D'Angelo, D., 1984, diversity of challenge is The Saltpetre-Moon Cave System: Burchett 's Cave (formerly Jarvie Pholeos, Wittenberg University Roark's Cave). This is a maze of Speleological Society, v. 4, no. criss-crossing, joint-controlled 1, 7-13 p. passages where solution has a1so occurred at different levels a1 ong Ferm, J. C., Horne, J. C., the bedding planes. Of particular Swinchatt, J. P., and Whaley, P. interest is a narrow passage W., 1971, deposi- called Avenue, where one tional environments in north- can see rimstone dams in excess of eastern Kentucky, jn Spring 1.5 m in height. Another Field Guide, Geological Society interesting section of the cave is of Kentucky, Kentucky Geological Heavenly Crawl, a low-ceilinged Survey, 30 p. passage where soda straw stalactites descend up to 46 cm. Horne, J. C., Ferm, J. C., 1977, Along the gorge of Tygart 's Carboniferous depositional envi- Creek from Olive Hill, Kentucky to ronments in the Pocahontas Basin the Car.ter-Greenup County line, of eastern Kentucky and southern caves of varying sizes can be West Virginia: Field Guide, found. Most are small, being less Departmen t of Geol og y , Un i v- than 91.4 m in length. Names given of South Carolina, 129 p. them seem to be sufficient description. Crawlsbad Caverns, McGrain. P., 1966, Geology of Disappointment Cave, H20 Carter and Cascade Caves Area: Cave, Impossible Cave, Kentucky Geological Survey, Moonshiner's Cave, Garbage Cave, Special Publication 12, 32 p. Cave, and Zig-zag- - Cave.. .a1 1 seem to offer little incentive to Miotke, F. D., and Palmer, A. N. visit except for those masochistic . 1972, Genetic re1ati onshi ps folk that seem to constantly be on between caves and landforms in the prowl for a "BIG" cave the Mammoth Cave National Park somewhere out there just waiting Area: Mammoth Cave, Kentucky, to be discovered. Mammoth Cave National Park,

Pf ef f er, N. , Madigan, T. J. , and Hobbs 111, H. H., 1981, Laurel Cave: Pholeos, Wittenberg Univ- ersity Speleologi cal Society, v. 2, no. 1. CAVES OF NORTHEASTERN KENTUCKY

Voigt, Keller, and Johnson, 1962, Ohio, Central Ohio Grotto, Caves of Carter County, Ken- National Spel eol ogi cal Society , tucky: COG Squeaks, Columbus, v. 5, no. 10, 31-39 p. Chapter 6 PINE MOUNTAIN KARST AND CAVES Joseph W. Saunders Agricultural Research Service Department of Agriculture 3207 Melody Lane East Lansing, Michigan 48912

Pine Mountain is a long, nar- ture represents the northwestern row, straight ridge extending 170 edge of the Pine Mountain km in a northeast-southwest direc- Overthrust Block and is bounded on tion in eastern Kentucky and the northeast by Russell Fork another 24 km in Tennessee (Fig. Fault near Elkhorn City (Kentucky) 1). This dominant topographic fed- and on the southwest by the

Figure 1. Pine Mountain, a large overthrust block, extends 105 miles in a northeast-southwest direction in eastern Kentucky forming the highest cave region in the State. PINE MOUNTAIN KARST AND CAVES

Jacksonboro Fault near Lafayette mountain where the dip into the (Tennessee). Pine Mountain encom- mountain is greater. Mountainside passes parts of Pike, Letcher, and ravines are much more Harlan, Bell and Whitley counties prominent on the outcrop slope in in Kentucky, and Campbell County the northern reaches of Pine in Tennessee. In Letcher and Pike Mountai n. counties, Kentucky shares Pine Moutain with Virginia, as the GEOLOGY state line lies at the mountain crest. Caves on Pine Mountain underlie Pine Mountain is a narrow- the northern (outcrop) slope of crested ridge of the Appalachian the mountain, the only position on Valley and Ridge type, and is the the mountain where limestones westernmost mountain in this part outcrop. Strata dip into the of the hppalachians. Due to the mountain at 20 to 40 degrees. The stratigraphy exposed by the thrust caves are found in the faulting, the coal so abundant and Mississippian Newman Limestone, important to the local economy in the local equivalent of the the Cumberland Plateau to the Greenbrier or Glen northwest, as well as the Dean-Girki n-Ste. Genevieve-St. Middlesboro Fasin to the south- Louis 1i mestones. &long Pine east, is absent on the outcrop Mountain the Newman Limestone is slope of Pine Mountain. between 122 to 183 m thick and is The elevation of Pine exposed about halfway up the Mountain's crest commonly ranges mountain. It is overlain by up to from 640 m to 700 m at the south- 215 m of sandstone, siltstone, and ern end to 792 m to 854 m at the shale of the northern end, with relief up to (Fig. 21, including the sandstone 488 m. The highest point on Pine mountain crest. Ninety meters of Mountain is approximately 997 m the Formation, mostly above sea level and located east siltstone and shale, lie of the community of Whitesburg. comf ortabl y be1ow the Newman The mountain is crossed by streams Limestone. The at only two places. The Cumberland (upper Devonian and Mississippian) River, the master stream in the is found beneath the Grainger and region, encompassing the southern up to 183 m of it extends portion of Pine Mountain, crosses downslope to the Fine Mountain the mountain at Pineville in a Fault near the foot of the narrow water gap at an elevation mountain. In places there are of 305 m. Near Jellico, 32 km to secondary faults in proximity to the west, Clear Fork crosses the the Pine mountain Fault (e.g., at mountain. To the east of Jenkins, Kentucky, and Newcomb, Fineville, the mountain runs for Tennessee), and in at least one 130 km unbroken by any surface such case a sliver of Newman stream, although minor shallow Limestone is exposed at the foot gaps in the crest occur at several of the mountain. ( Interesting1y, places. Interstate 75 south of there appears to be a subterranean Jellico climbs the outcrop slope meander cutoff of Elk Fork in such of Pine Mountain before traveling a situation near Hells Point along the crest for nearly 16 km. Ridge, Tennessee.) The Newman The re1ati vel y homogeneous topo- itself is divided into a larger graphy of Pine Mountain is remark- lower member of limestone and a able on its northern upper slope smaller upper member consising of for its relative homogeneity, with shale, 1i mestone, and sandstone. contours running nearly straight Most, if not all, the known caves and parallel for kilometers. This are in the lower member, and most homogeneity is especi a1 1y true appear to be in the low, commonly along the southern reaches of the dolomitic portions of the lower 88 GEOLOGY

The water gaps through the mountain at ~inevilleand Jellico P are associated with faults in the z 2 Upper overthrust block. The gap at a 4 Pineville is on the Rocky Face zz I= c member Fault and is related in origin to a a -.-0, the Cumberland Gap 16 km to the 0- south (Rich, 1933; Froelich and q ,a, Tazel aar , 1974) . Qn array of I>',E faults exists on Pine Mountain acn cL just east of the gap at Pinevi lle. 00 Ez Lower The mountainside over1ying the ~z ak member Newman Limestone is steep, and 2; ravines and hollows cut into the z limestone commonly have braided rubble streambeds. Solution dolines are few in number. Because of the relative1y short distance Upper from the top of the mountain to member the base exposures of the a, limestone, usually 305 to 450 m, z c t stream flow in these small hollows a and ravines is intermittent. There I a 0 n is recognized potential for loss n EZ of some or all such intermittent I 3 a, Lower cn flow into the limestone under the cn a, E member rubble. In the case of Payne Gap I 2 -1 cn Cavern (Figs. 3 and 4), the cn overlying ravine has intersected I the cave passage. This z intersection has resulted in an entrance that takes some of the intermittent stream flow coming down the ravine. Several pits are Grainger known on the mountain, including Colehole 160 to 73 m deep) and a z Formation vertical entrance to Linef ork 2 4 Cave. Springs are fairly common in a z the limestone, presumably near the z n base. They range from small ones f 3i with no cave opening to larger z Chattanooga springs with base flow of about 0; 14.2 l/s (liters per second). The >cn Shale limestone springs on Pine Mountain W- tend to be larger and cleaner P I compared with springs in the surrounding area, and commonly are employed as domestic or community Figure 2. Stratigraphic profile water sources for people 1iving at of the outcrops exposed on Pine the foot of the mountain. Caves Mountai.n . are commonly found at or near springs on Pine Mountain, and in some instances access to water member, a1thouh lack of exposure supply caves has been difficult. of the basal contact and paucity Numerous large vertical faces of of detailed stratigraphic study limestone are found along the within the Newman preclude an mountain, commonly above the authoritative statement. springs. PINE MOUNTAIN KARST AND CAVES

Cross Section AB A- \/ PROFILE

Cross Section CDE

,',I LE PAYNE GAP CAVERN ' / Letcher County, Kentucky

C.R.G. Grade 4 Map Completed in 1974 o SCALE 30 Feet by Gary Jessey, Keith Ortiz, and Joe Saunders - T.H.C. 455 Feet Figure 3. Payne Gap Cavern, Letcher County, Kentucky--a three-level cave containing a base-level streamway.

CAVES ceiling and floor in the upstream direction (Fig. 6). Although the There are nine mapped caves on load reduction bank of sand is Pine Mountain, and 15 to 20 known often found at sumps, it might but unmapped caves. The longest also have been deposited in a known cave is Linef ork Cave in flood event. An examination of the Kentucky, where a survey in relation of Linefork Cave's progress has mapped 2,440 m along passage to the overlying the strike (Fig. 5>,corresponding topography reveals that the to a surface distance of 1,830 m. downstream (western) end of the The cave's main entrance is near a cave is found where the hillside major spring, and five sumps in is steepest. If we assume that the the main stream are bypassed dip into the hillside is through higher level passages. A relatively uniform in magnitude, sixth has presently halted then the spring below the main exploration. There are indications entrance at the western end that there has been a terminal represents the lowest elevation in sump in this area of the cave for the area that the limestone is some time, namely the descending exposed. One might expect drainage CAVES

PAYNE GAP WATER CAVE Letcher County, Kentucky C.R.G Grade 4 Map Completed in 1974 Gary Jessey, Keith Ortiz, and Joe Saunders

JENKINS WEST 7.5' QUAD.

T., +7 SCALE Newman Limestone 0 25 50 Feet - T.H.C. 403 feet

Figure 4. Payne Gap Water Cave, Letcher County, Kentucky--an example of an Appalachian-strike single-passage cave.

within the limestone to be Pine Mountain in Kentucky, all discharged at this lowest possible around 152 m in length. Two of outlet, although the alternative these caves are in the explanaton of the spring branch area near Jenkins and have been being responsible for the hydro1ogi callY connected with a topographic situation must also be f 1uorescei n trace (Saunders, weighed. Linefork Cave has at 1974). Payne Gap Cavern has three times in the past, been referred distinct levels including a base to as Water Cave. The present level streamway, whereas Payne Gap survey project in Linefork Cave is Water Cave downstream is in some reserving the name Water Cave for ways the stereotypical Appalachian the 45-m-long spring cave located strike single passage cave (Figs. immediately below the main 3 and 4). Icebox Cave at the water Linef ork entrance and gap at Pineville may be associated corresponding to the downstream with faulting in the mountain side of sump no. 1 in Linef ork there. Cave. The Colehole in Letcher County In addition to Linefork Cave, is the deepest known pit on Pine there are five caves mapped on Mountain. The entrance to the pit PINE MOUNTAIN KARST AND CAVES 9 1

Figure 5. Linefork Cave, the longest mapped cave on Pine Mountain.

"Ieo sump 0 SCALEI 50I Feet I - ...... _ sand"...,

Figure 6. Exploration in Linefork Cave has been stopped by a terminal sump formed from a descending ceiling in the upstream direction. LONG MOUNTAIN SETTING is a hole 1.5 m in diameter. The influence of the geological geologic structure of the mountain structure on passage trend is is reflected in the pit by ledges fairly consistent as we1 1. Thus, and offset vertical drops. From passages usual 1y f 011 ow the ground zero at the entrance there mountainside, near1y para1lel to is a sheer vertical drop of 27 to both the outcrop strip and the 30 m onto a steeply sloping ledge. mountain crest. Dip tubes (i.e., The largest area in the pit is at steeply sloping passages draining this ledge, where the shaft may be from the surf ace in a downdip 12 to 15 m in diameter. Below the direction at right angles into the first ledge is another drop onto a strike passages) have not been second ledge. A third drop reen by the author in Pine immediate1y f olows, descending to Mountain caves. They are seen in a breakdown slope at the bottom of caves in other long mountains in the shaft. Total depth below the the Appalachians, where they may entrance has not been accurately be associated more with greater determined but is between 61 to 73 mountainside do1 ine +requency. m. Horizontal gallery passage in LONG-MOUNTAIN SETTING Pine Mountain caves is relatively irregular in cross section. Pocket Pine Mountain is one of the and features suggest that long mountains in the central phreatic development played a very Appalachians of the eastern United significant 1-01 e in passage States that were discussed by formation, perhaps similar to the Saunders and others (1977) as the role envisioned by Bretz (1942) in setting for a genre of karst, Cudjo's Cave in Cumber1and caves, and subterranean hydrology. Mountain 16 km south of Pineville, Long- mountain karst occurs where under simi 1 ar structural and limestone or dolomite units crop stratigraphic conditions. out in continuous strips along one A characteristic feature of side of linear sandstone or Pine Mountain caves ik, that many conglomerate capped mountains, of the passage cross sections paralleling the strike and the display the effect of the steep mountain crest, and dipping into dip. Ceilings, ledges, and the mountain. Whereas the sometimes floors often slope into stratigraphy may differ from the mountain (Fig. 7a-b). The mountain to mountain, the

Figure 7. Cross sections of passages in Pine Eiountaic caves show the effect of the steep dip of the formations. PINE MOUNTAIN KARST AND CAVES

structural setting is similar and band along the steep mountain and fairly regular, with beds dipping continues on down the mountainside 5 to 40 degrees, depending on the surface in the numerous ravines particular area. The re1ativel y and hollows. It has not been uniform relationship between ascertained how many sites in the stratigraphy, structure, and 1i mestone swal 1ow such topography along each mountain intermittent stream flow either should permit generalizatons about partial1 y or who1 1y (fewer are the development of caves and known on Fine Mountain than on subsurface drainage from the other long mountains) , nor to what number of examples that arose extent f 1ood pulses are under seemi ngly simi 1 ar experienced in the downstream circumstances. In the case of Pine portions of the subsurf ace Mountain, the cave and spring drainage systems. It is presumed record is so incomplete that that the range of discharge generalizations based on Pine volumes is re1ati vel y narrow for Mountain examples alone are risky. Fine Mountain springs if much of Nevertheless, one can pose the wetter weather flow goes questions drawn from other long uncaptured as it flows down the mountains and try to answer these mountainside. A related matter is with available data. the degree to which water levels Drainage basins of Pine in the caves rise during extreme Mountain springs are long and precipitation events, and the narrow. With up to 213 m of extent to which cavers might be clastic sediment lying between the endangered. In the judgement of Newman Limestone and the crest of this author, both are relative1y the mountain, the widths of the minor. drainage basins are between 245 and 460 m. Distance between WATER GAPS springs on Pine Mountain appears from incomplete field work to be Perhaps the most interesting up to 5 km. Thus drainage basins question pertaining to Pine . are up to 4,575 m long and 460 m Mountain caves and subsurface wide. They are oriented along the drainage is the comparison between strike, and thus are concordant gap and mountainside caves. Most with -most of the adjacent major of the springs and caves on Pine surf ace drainage. Drainage under Mountain lie on the mountainside the mountainside to a spring from several hundred meters above one direction on the mountainside present-day base level surface only is the rule in the limestones streams. This is because the in the long mountains. On Pine limestone is exposed on the Mountain, information is available mountainside we1 1 above the for on1y three mountainside surface streams. Mountainside subterranean drainage systems. Twa springs can be viewed as spillover appear to drain from one direction points for water flowing within only (Linefork Cave and Payne Gap the limestone, most of which in Water Cave). Rose Sping Cave in the mountain lies below the Tennessee is the exception, as it elevation of the spring. As drains from two directons (Fig. spillover points, most springs can 8). There is evidence from the be expected to represent the western part of that cave of an 1owest elevation exposures o+ abandoned outlet, suggesting the limestone along the mountainside. fusion of two drainage systems Due to the steep dip into the into one at some time in the past. mountain, incised topographic Drainage for these basins features in the limestone outcrop cannot be considered fully captive such as deep ravines and hollows because during wetter weather some represent the lowest elevations on runoff 'passes over the limestone the mountainside where discharge WATER GAPS

ROSE SPRING CAVE VERY LOW CRAWL OVER SMALL RUBBLE: BREEZE Campbell County, Tennessee C.R.G. Grade 4 Map Completed in 1976 by Jim Borden and Joe Saunders

SCALE 0 50 Feet

36O 34'1 2' N., 84O 06'25" W. JELLICO EAST 7.5' QUAD Elevation Approx. 1500' T.H.C. 419 feet Lower Newman Limestone

UNENTERED ROOM IN UNSTABLE BREAKDOWN: BREEZE Figure 8. Rose Spring Cave in Campbell County, Tennessee, drains from two directions, an unusual occurrence on Pine Mountain. from the limestone is possible. the lowest mountainside 1i mestone Extreme examples of this should be exposures a mile or more away found at water gaps where major (Figs. 9 and 10). hlthough this base level surf ace streams cut exposure at river level would be through the mountain. Here the expected to allow gap springs to limestone is found at river level compete very we1 1 with several hundred feet 1ower than mountainside springs for drainage PINE MOUNTAIN KARST AND CAVES 9 5

Figure 9. Water-gap exposure of Figure 10. Most springs and limestone at river base level. caves lie on the mountainside Expected cave for~ationat the several hundred feet above gap spring did not naterialize. present base level of surface streams . area, and to allow discharge sites kilometer east of the gap there is for the development of long and a spring on the mountainside deep caves, there is little nearly 92 m higher in elevation evidence for this at the water than the river gap. No cave was gaps at Pinevi 11e and Jelli co. noted there, and no dye tracing There are three small caves and a was attempted to determine the small spring known on the east flow direction to the spring. side of the gap at Pineville. It is not clear how to account While no spring was noted on the for the apparent absence of major western side of the gap at springs or caves at the Pine Pineville, there is a high water Mountain water gaps. One outlet in which the owner will not suggestion has been that in case5 permit exploration. Nothing was where the water gap appears to owe found on the eastern side of the its existence to a fault running gap at Jellico, and a very minor through the mountain, such as at spring was located on the western Pineville and Jellico, joints and side. Hose Spring Cave on the bedding partings otherwise mountainside, 2.5 km west of the available for the development of gap at Jellico and more tha 92 m solution openings and conduits higher in elevation, drains from have been adversely affected the directon of the gap. during development of the fracture CONCLUSION zone. fin a1ternative explanation counterparts in all other areas of involves recent or rapid changes Kentucky. They are unique because in regional base level inasmuch as Pine Mountain lies in the small this might relate to time portion of the State considered to available for gap springs to be in the Appalachian Valley and capture drainage in a headward Ridge physiographic province, with direction at the expense of its charactristic topography and mountinside springs. One has only geologic structure. Formal to look 16 km south of'Pineville exploration and study of the caves to Cumberland Gap to see the has been limited. It is recognized potential for cave development at that the potential for new water gaps. There, with discoveries exist, including pits stratigraphy and structure similar and caves of moderate length to Pine Mountain, Cumberland Gap ti. e. , approximately 2 km) . Saltpetre Cave has a length of several miles and a depth of several hundred feet. A large spring rises at the gap and has been traced by Dr. James F. Buinlan from 15 km away on a REFERENCES mountainside. One difference in the situation there is that Bretz , J. H. , 1942, Vadose and Cumberland Gap is an abandoned phreatic features of limestone water gap now several hundred caverns: Journal of Geology, v. meters above the major surf ace 50, p. 675-811. streams. The abandonment is thought to have involved the Froel ich, A. J. and Tazelaar, piracy of a stream in the J. F., 1974, Geologic map of the Middlesboro Basin west of the gap Pineville Quadrangle: U. S. Geo- from the drainage logical Survey Geologic Buadran- to that of the Cumberland River gle Map 68- 1129. (Rich, 1933). A possible explanation for the greater cave Rich, J. L., 1933, Physiography and hydrosystem development at and structure at Cumberland Gap: Cumberland Gap compared to the Geological Society of America gaps at Pinevil le and Jellico is Bulletin, v. 44, p. 1219-1236. that the postulated water gap stabilized the elevation of the Saunders, J. W., 1974, Payne Gap spring, whereas at Pinevi 1le and caves, Pine Mountain, Kentucky: Jellico the spring outlet COG Squeaks, Central Ohio Grotto elevations have been continually National Speleological Society, dropping as the Cumberland River v. 17, no. 12, p. 125. and Clear Fork downcut, dropping at a rate too fast to permit development of solution channels Saunders, J. W., Medville, D. M., capable of headward growth. and Koerschner, W. F., 1977. Karst drainage patterns of the CONCLUSIONS long mountains of the eastern United States: Seventh Interna- Pine Mountain karst and caves tional Congress of Spel eology, are unique with regard to their p. 375-376. Chapter 7 THE MAMMOTH CAVE REGION AND PENNYROYAL PLATEAU Arthur N. Palmer Department of Earth Sciences State University of New York Oneonta, New York 13820 The Mammoth Cave Region is by The cavernous limestone is far the best known karst area in underlain by impure, poorly the United States and is the one karsted 1i mestone, a1so of Missi s- most ~0mm6nlycited in articles on sippian age, and is oveqlain by cave origin. Since it has been sandstone and shale of Mississip- described in considerable detail pian and Pennsylvanian age, which in previous works, only a brief contain thin units of limestone summary of the specific karst general1y no more than 10 m thick. landscapes and cave systems will The insoluble rocks form hilly, be given here. Instead, the main dissected plateaus in the center purpose of this chapter is to pro- of the 11 1inois Basin, but around vide insight into the origin of the edges of the basin they have these features and to determine been removed by erosion, exposing what makes the area unique. the limestone at the surface (Fig. 3). The limestone forms a broad, Only the area in and around low-relief karst plain, typified Mammoth Cave National Park is con- by sinkholes and sinking streams, sidered in detail in this chapter called the Pennyroyal Plateau. The (Fig. I).Except to the northwest, deeply dissected perimeter of the the karst of this region extends a hi11 y region consists of 1imestone great distance in all directions, ridges chapped by insoluble rocks diminishing somewhat in intensity. and separated from one another by Although these surrounding areas steep-wall ed val 1eys. Many of are not specifically treated here, these are perched karst valleys at the general concepts apply to them the general level of the as we1 1. Pennyroyal Plateau (Fig. 4). This borderland is called the Chester REGIONAL SETTING Upland, and the sharp, irregular, 60-m-slope that separates it from The area covered by this chap- the Pennyroyal Plateau is called ter is located in west-central the Chester Escarpment (Fig . 5) . Kentucky at the southeastern edge The Pennyroyal is subdivided into of the and is two areas: a sinkhole plain, which underlain by limestone of Hissis- is 1ocated on re1ati vel y pure sippian age (Fig. 2). These rocks limestones bordering the Chester dip gently northwestward toward Upland, and the Glasgow Upland, the center of the basin at 5 to 10 which is an area of large1y sur- m/km. The total thickness ofcaver- face drainage on silty and shal y nous limestone in the region is limestone lower in the section. on1y a few hundred meters--rather Many of the streams on the Glasgow thin in comparison to most other Upland sink into karst depressions karst regions of the world. HOW- where they flow onto the purer ever, its small angle of dip 1i mestone. allows it to be exposed over a wide area, so that extensive karst The Chester Upland and drainage basins of up to 500 Pennyroyal Plateau form a km" have developed. crescent-shaped band of karst that 98 REGIONAL SETTING

Figure 1. Physiographic map of Kentucky, showing the location of the Mamrno th Cave Region. extends around the southern and groundwater. The largest of these eastern edge of the Illinois Basin are the and its from southern I1linois, through tributary, the Barren River, at a western Kentucky, into southern local altitude of 130 m. Tributary Indiana. The karst area diminishes to these are numerous karst basins northward because of decreasing with main1y subsurf ace drainage. limestone thickness and an These basins have been delineated overburden of glaci a1 deposits. over the past 10 years by James The broad-scal e configuration of Buinl an of the Up1and5 Research this karst area is determined Laboratory at Mammoth Cave with primarily by the form of the the aid of dye traces and Illinois Basin, but its local measurements of static water level details are controlled by broad in we1 1s. These basins are shown open f 01ds. in generalized form in Figure 6. The region is crossed by The largest karst basins lie several entrenched rivers that either entirely within the serve as outlets for karst Pennyroyal or have headwaters in THE MAMMOTH CAVE REGION AND PENNYROYAL PLATEAU 99

Figure 2. Stratigraphy of the Mammoth Cave area and its relationship to the surrounding landscape. 100 REGIONAL SETTING

Figure 3. Generalized profile through Mammoth Cave. A = caprock of interbedded sandstone, shale, and limestone (Chesterian Series); B = Girkin Formation; C = Ste. Genevieve Limestone; D = St. Louis Lime- stone; E = impure limestone and detrital rocks; F = upper cave levels of late Tertiary age; G, H = typical levels of Quaternary age (partial flooding of the lowest is due mostly to late Pleistocene alluviation of the Green River).

Figure 4. A typical sinkhole-floored karst valley in the Chester Upland. The flanking ridges are capped by insoluble rocks of Chester- ian age. THE MAMMOTH CAVE REGION AND PENNYROYAL PLATEAU

Figure 5. View from the Chester Upland (at left) into the Pennyroyal Plateau near Mammoth Cave. the Pennyroyal and drain to the on the Green River. & few of its Green River through the Chester upstream branches can be seen as Up1 and. minor sinking streams on the The largest karst springs of Glasgow Upland. Other tributaries the area are fed by passages 10 to include a lengthy system of river 15 m below present river level, passages in the southern part of rising upward through openings in Mammoth Cave. The main underground a1 luvia1 sediment (Watson, 1966; stream appears at the bottom of Hess, 1976). The area lies more Mi 11 Hole, an impressive karst than 80 km south of the farthest window at the southern edge o+ the 1imi t of Pleistocene continental Chester Upland, and can also be glaciers, so it was spared the seen in caves at the base of Cedar direct effect of glaciation. But Sink, a huge collapse sinkhole in the Green River is a tributary of the bottom of Wool sey Val1 ey. the Ohio River, whose base level However, these scattered fragments was greatly affected by give only a crude idea of the glaciation, and the 15-m-thick exact pattern and nature of the alluvial deposits of the Mammoth conduits that present1 y 1i e be1 ow Cave area are the result of base level. aggradation of the Ohio River during the Wisconsinan glacial STRATIGRAPHY advance. The major flow routes delineated by dye traces are The largest caves of the region accessible on1y in a few scattered are developed within a continuous cave passages. Most of the large sequence of Missi ssi ppi an stream conduits are perennially 1 imestone formations having a flooded because of alluviation of total thickness of approximately the river valleys. For example, 150 m (Fig. 2). These include, one of the largest karst basins is from bottom to top, the St. Louis the one that feeds Turnhole Spring Li mestone, Ste. Genevi eve 102 STRATIGRAPHY

CHESTER UPLAND

Sink hole Plain

2 Crump Spring ll Mammoth 3Fisher Ridge 12 Northtown

14 Proctor

#= Generol ized Cove Maps

---/- .--- Glasgow Upland

Figure 6. Location map of caves and subsurface drainage basins in the Mammoth Cave area (modified from Quinlan and Ray, 1981).

Limestone, and Girkin Formati on. been removed by erosion, but they These rocks are laterally very reappear in eastern Kentucky and extensive. The St. Louis and Tennessee and in northern Olabama underlying rocks maintain their and Georgia, where the bottom-most identity over great distances insoluble unit (Hartsel le throughout the east-central United Sandstone) is correlative with the States. The Ste. Genevieve and of the Girkin are less uniform in Mammoth Cave area. The major character, for the upper part of cavernous limestone formations are the sequence becomes interbedded described in the following with insoluble detrital rocks paragraphs. toward the northwest in Illinois The St. Louis Limestone and Indiana. The effective top of consists of 50 to 60 m of shaly, the cavernous zone is near the silty, and cherty limestone and base of the Girkin (Paoli Member) do1 omi te. Karst devel opment is in Indiana and as far down as the greatest in the upper half of the middle of the Ste. Genevieve (Spar formation, where shaly and silty Mountain Member) in I11 inois. On1 y beds are fewer. At depth below the small caves occur in the thin surface. the St. Louis contains limestone units sandwiched between extensive gypsum beds, but they the insoluble rocks above these are usual 1y leached out by h~rizons.To the southeast, the groundwater long before they have Ste. Genevieve and Girkin have a chance to be exposed in caves or THE MAMMOTH CAVE REGION AND PENNYROYAL PLATEAU at the surface. Bedded and sediment carried from the adjacent lenticular chert occurs throughout land areas created a broad delta most of the formation and is (Michigan River Delta) that particularly abundant near the extended progressively across the top. The St. Louis is underlain by central part of the sea (Swann, the Salem and Harrodsburg 1964). The thin beds of impure formations, which are rather 1imestone in the Girkin and upper impure shaly limestones with a Ste. Genevieve are the early combined thickness of 30 to 50 m. precursors of this invasion of Karst development is rather poor detrital sediment, and the thick in these rocks, and caves are sandstone and shale of the Chester mainly small and perched on Series represent the advance of insoluble beds. The Glasgow Upland the delta into the area. After a is developed on the lower St. period of f luvial erosion at the Louis and the Salem-Harrodsburg, end of the Mississippian Period, and the sinkhole plain is the region was blanketed again by developed on the purer upper St. Pennsylvanian conglomerate, Louis and to a lesser degree on sandstone, and shale, which are the Ste. Geqevieve. exposed in the central parts of The Ste. Genevieve Limestone is the Illinois Basin. The about 35 m thick. It is probably unconformity at the base of the the most cavernous limestone unit Pennsylvanian rocks has a maximum in the region and contains most of relief of nearly 100 m. There is the Mammoth Cave System and nearby 1ittle evidence for a caves. It consists of interbedded Mississippian paleokarst , so 1imestone and dolomite in the prevalent in the western United lower ha1f , and 1imestone States, because most of the interspersed with thin beds of pre-Pennsyl vani an erosion was incompetent silty beds in the confined to the insoluble rocks upper half. Nodular chert is capping the limestone. common near the top and in the Figure 7 shows the middle of the formation. stratigraphic changes in the The Girkin Formation, 25 to 40 Girkin Formation as it is traced m thick, consists of limestone northward and westward from the with thin interbedded shale and Mammoth Cave area. In both siltstone. It is a favorable rock directions the formation becomes for cave development where progressively more partitioned by geomorphic- conditions are thick shale and sandstone units, suitable. The insoluble beds are breaking the hydrologic continuity generally less than 1 m thick and of the limestone. This trend might do not significantly interrupt the help explain why the Mammoth Cave continuity of cave development in area is so favorable to cave the area. The Girkin is overlain development . However, very few by the , the passages in Mammoth Cave occupy basal sandstone of the chief 1y the upper beds of the Girkin, so insoluble caprock that forms the this explanation alone is ridge crests of the Chester inadequate. Wore important1y, the Upland. Thin limestone units topographic re1ief is greatest in between the insoluble rocks the Mammoth Cave area, with 90 to contain perched karst drainage, 140 m of continuous limestone and springs at their basal exposed above base level. This contacts provide the water supply figure dimi ni shes gradual 1y to Mammoth Cave National Park. northward to about 65 m and These limestone formations were westward to less than 45 m. deposited in a shallow continental Dissection of the Mammoth Cave sea that extended across most of Region by the Green and Barren southern North America during the rivers was even deeper in the Mississippian Period. Detrital past, because the valleys are now 104 MAJOR 'CAVE SYSTEMS

Figure 7. Lateral variation in the Girkin Formation and correlative rock units in western Kentucky. The Girkin consists of uninterrupted limestone in the Mammoth Cave area but is partitioned into separate formations by insoluble rocks to the north and west. (Correlation is based mainly on geologic maps by the U. S. Geological Survey.) floored by roughly 15 m of late huge system. Generalized maps of Pleistocene sediment. Furthermore, the major caves are shown in the position of these deeply Figure 3. entrenched rivers with respect to filthough distinct karst the broad extent of exposed drainage basins have been mapped limestone is ideally suited to the in the region (Quinlan and Rowe, devel opment of underground 1977; Quinlan and Ray, 1981), many drainage. of the divides lose their identify during high +low as numerous MAJOR CAVE SYSTEMS overflow routes become active. Because the underground divides The caves of the Mammoth Cave also shift with time, dry passages area are notable not so much for connect several caves that are now the size of their passages, which 1ocated in separate basins. is matched or exceeded by many caves elsewhere, but for their Caves of the Chester Upland unusual length and interconnectivity (Fig. 6). Caves Chief among the caves of the once thought to be entirely region is of course the Mammoth independent have been linked with Cave System, which now includes such regularity in recent years its connected neighbors Flint that it might seem only a matter Ridge, Proctor-Morrison, and of time before every cave in the Roppel caves, with a total length region is interconnected. Although of 490 km. Mammoth Cave is the such a feat will probably never be world's longest cave by a factor achieved, it is still appropriate of more than three. Its sheer size from the genetic standpoint to is what has drawn the attention of consider nearly all the caves of explorers, writers, and scientists this region to be part of a single for nearly 200 years. The system THE MAMMOTH CAVE REGION AND PENNYROYAL PLATEAU 105 ex tends under at 1east five to heights of as much as 20 m different ridges of the Chester (Fig. 10). Upland and beneath the karst valleys that separate them. It is The system is being explored surrounded by many caves that are and mapped by the Cave Research enormous in their own right, some Foundation (CRF), which was of which are hydrologically or organized in 1957 by the explorers genetically related to Mammoth. and commerci a1 operators of Floyd Mammoth Cave has been formed by Collin's Crystal Cave. The drainage from not only the Chester earliest history of exploration by Upland, but also from adjacent this group is described in THE areas of the Pennyroyal Plateau. CAVES BEYOND by Lawrence and The water emerges at several large Frucker (1955). By 1961, CRF springs along the Green River. explorers had linked most of the Mammoth Cave and all others in the caves of Flint Ridge into a single Chester Upland are typified by system, which became the world's long tubes and canyon passages at longest in 1967. In 1972, a many 1eve1 s, with vertical shafts connection was found beneath concentrated around the perimeters Houchins Val ley to Mammoth Cave, of ridges (Figs. 8 and 9). creating a single cave 230 km Recharge to the caves comes both long. The explorational history of from nearby parts of the the Mammoth Cave System up to this Pennyroyal and from karst valleys point is told in THE LONGEST CAVE in the upland. Except for active by Brucker and Watson ( 1976) . canyons, the upper-level passages Since then the cave has been are rather dry, particulrly those connected to Proctor Cave, beneath the nearly impermeable Morrison Cave, and Rappel Cave by caprock. Lower levels are still an extensive series of river active and are subject to passages in th St. Louis backflooding from the Green River Limestone.

Figure 8. Turner Avenue, in the Flint Ridge section of the Mammoth Cave System, is a shallow-phreatic tube of Quaternary age now situated at an altitude of 168 m.. 106 CAVES OF THE CHESTER UPLAND

Cave. It was discovered in 1981 and has been explored and mapped to about 47 km by the Detroit Urban Grotto of the National Speleologi cal Society. A1though some of the more promising passages have led to sumps, the potential for additional discovery is great, for streams in the cave receive water from as far away as Cave City. Crump Spring Cave (19 km long) is located in the northern arm of Fisher Ridge. Since its discovery in 1965 it has been explored by groups from a large geographic area, most recently under the direct i on of Joseph Saunders of Lansing, Michigan. The cave is a complex of canyons and vertical shafts with active and abandoned drains on many levels. Whigpistle Cave, to the southwest of Mammoth Cave, was discovered in 1978 by personnel of the Uplands Research Laboratory and has since been mapped to 32.5 km. It is a dangerously wet cave that drains several local karst valleys. Its largest passage, which ends in breakdown at Woolsey Val ley southwest of Joppa Ridge, Figure 9. View downward into a seems to be an upstream fragment 200 m deep vertical shaft in of part of New Discovery at the Mammoth Cave. western edge of Mammoth Cave. Roppel Cave, in the northern A11 of the caves described end of Toohey Ridge east of above 1ie south of the Green Mammoth Cave, was discovered in River. Directly north of the 1976 and has been explored tn more river, as we1 1 as farther than 80 km by the Central Kentucky downstream toward the west, caves Karst Coalition (CKKC). In are cornparati vel y small because character it is similar to Mammoth the exposure of limestone is Cave, although many passages show limited by an increasingly thick evidence of flow reversals and clastic caprack. piracy. Most of the larger To the south and southeast of passages seem to be truncated Mammoth Cave National Park are upstream fragments of passages in many residual limestone knobs Mammoth, although the exact presently or formerly capped by correlation is not yet clear. The outliers of sandstone (Fig. 3). An single connection with Mammoth example is Bald Knob, which Cave was discovered in 1983 by a contains James and Coach Caves. combined CKKC-CRF team through a These are complex, long river passage in the St. three-dimensional caves consisting Louis Limestone (Forden and of numerous interconnecting Crecelius, 1984). canyons and shafts, and with Fisher Ridge Cave is located a high-level network mazes directly short distance northeast of Roppel under1ying the sandstone cap. THE MAMMOTH CAVE REGION AND PENNYROYAL PLATEAU 107

Figure 10. Logsdon River, part of an extensive dendritic system of active passages in the southern end of Mammoth Cave. The upstream end of this passage connects with Roppel Cave.

Caves of the Pennyroyal Plateau creamery was so severe that the operation had to shut down. In The extent of explorable caves 1975, 15,000 liters of gasoline in the sinkhole plain became were lost in the system from a apparent only in the early 1970's. leaky tank, and gasoline was Olthough sinkholes are numerous, smelled in local . The most are clogged with soi 1, water from Hidden River Cave col 1 apse materia1 , and assorted divides downstream into numerous debris. Caves are wet and distributaries and discharges at dangerous due to the possibility 46 springs over an 8-km reach of of rapid flooding. Because the the Green River. The cave is part sinkhole plain is relatively flat of an enormous karst drainage and has a fertile soil, it is more basin of 500 km2. densely populated than the ridges Hicks Cave is a major part of and groundwater pol1 uti on is this distributary system and takes widespread. Despi te these much of the overflow from the drawbacks, recent years have seen Hidden River System. Hicks Cave is a great increase in the number and being explored and mapped by length of explor~dcaves in the Uplands Research Laboratory cavers sinkhole plain. Most of these are and now has 31.4 km of surveyed developed in the St. Louis passages, most of which are part1y Limestone. filled with water nearly at the Hidden River Cave, in the town present river level. of Horse Cave, was commerci a1 ii ed Gradys Cave is located in the in 1916, but by 1943 the pollution lower St. Louis Limestone in the from domestic sources and a nearby far eastern portion af the KARST AREAS NOR.TH OF MAMMOTH CAVE sinkhole plain. It was first Although the karst and cave explored in the middle 196QSs, but development in this northern area has since been pushed farther and is similar in many ways to the mapped to a length of 19.3 km by a Mammoth Cave Region, there are group under the direction of several features that make it Joseph Saunders. Most of the cave unique. The boundary between the consists of river passages Pennyroyal Plateau and the Chester situatea at base 1eve1 , with Upland is not so clearly defined numerous overflow routes. as it is near Mammoth Cave, and Parker Cave, near Park City. is the largest cave system, the located near the headwater area of Sinking Creek System, extends both the Turnhole Spring drainage under ridges and sinkhole plain. basin. It is distinctive in having The longest single cave in the five independent para1lel stream system is Fig Bat Cave, with 20.2 passages connected by a single km of mapped passages. The cave is upgraded overflow route that part of a karst of transmits water in various 376 km2 that discharges to directions depending on the Boiling Spring, which has an pattern of flood pulses in the estimated peak flow of more than various stream passages. The cave 50 mS per second (George, was explored by Uplands Research 1976). Some caves in the area are Laboratory personnel and mapped to very shallow, and there are a total length of 9 km. extensive cave pasages that have To the southwest of the area been partially unroofed by erosion shown in Figure 6 is a vast area and exposed as open trenches. of sinkhole karst draining These passages generally terminate westward to the Barren River. The upstream and downstream in remnant largest karst basin is that caves that have not yet been feeding Graham Spring, which unroaf ed . drains S10 km2. The city of Bowling Green is located over a RELATIONSHIP OF similar karst system. The city has KARST AND CAVES occasional problem5 with TO THE GEOLOGICAL SETTING ndustr ia1 and domest ic i Stratigraphy exerts the primary , drainage control over the karst landscape, problems, and backflooding of as shown in Figure 2. The clearest sinkholes during wet periods. The expression is the contrast between work of the Center for Cave and the Chester Upland and Pennyroyal Karst Studies at Western Kentucky Plateau, determined by the Universiy is directed main1y presence or absence 04 insoluble toward understanding and Chesterian rocks. On a more subtle alleviating these problems. scale, the relative1y small vertical permeability of the Karst Areas North Salem, Harrodsburg, and lower St. of Mammoth Cave Louis inhibits karst and promotes The Chester Upland and surf ace drainage. The areas of Pennyroyal Plateau extend greatest sinkhole and cave northward into southern Indiana development in the Pennyroyal (Fig. 1) . A1 thouh topograhic Plateau are formed by the upper relief diminishes slightly in this St. Louis beds, which are direction and the Girkin Formation relatively pure except for their becomes partitioned by detrital chert content. Deep-seated removal formations, the region between of gypsum by solution may account Mammoth Cave and the Ohio River for some of the fracturing and hosts some of the finest karst and small-scale distortion o+ these caves in the country. The reader beds, which in turn may foster the is referred to George (1976) for a extensive sinkhole development. more complete description. The even purer carbonate rocks of THE MAMMOTH CAVE REGION AND PENNYROYAL PLATEAU the Ste. Genevieve, on the other successive 1ateral drains within hand, develop broad, shallow each bed. sinkholes that are fewer than The nature of any given cave those in the St. Louis. Whether passage depends strongly on this distribution represents a whether it originated in the true stratigraphic control is vadose zone or in the phreatic obscured by the fact that the area zone. Most passages that form in of the sinkhole plain where the the vadose zone (canyons and Ste. Genevieve is exposed lie at perched tubes) have an almost the foot of the Chester perfectly consistent downdip Escarpment, where insoluble . orientation. Many passages of col luvium accumulates from the phreatic origin trend nearly up1and and f i11 s karst parallel to the local strike, depressi ons . because in these prominently Caves of the region are bedded rocks the most efficient af fected rather uniformly by the path for phreatic water is usually different strata, and the major at shallow depth at or just below differences in cave type are the water table. Fractures tend to imposed more by the local become tighter and fewer with hydrologic and geomorphic setting. depth, and so few large ones cut The thin but prominent bedding and across the bedding that deep flow numerous small joints cause the along one bed is usual 1y not able cave to be more sinuous and to pass upward into overlying concordant to the strata than in beds. Nevertheless, some passages most other karst regions (Deike, are discordant to the bedding and 1967; Palmer, 1977). The basic show evidence of phreatic water pattern of passages is dendritic, that rose in the downflow but this pattern is usual1y direction. The discordance is obscured by the numerous overflow almost invariably in a competent, routes in the Pennyroyal caves and thick-bedded limestone, such as by the many superimposed levels that of the Girkin Formation, the and diversions to progressive1y middle and lowest beds of the Ste. lower levels of the Chester Upland Genevieve, or certain massive beds caves. in the upper St. Louis. Echo River The prominent bedding and in Mammoth Cave shows several such general lack of tectonic jumps from one bed to another. disturbance has allowed cave The sinuosity of both phreatic passages to develop very clearly and vadose passages is strongly defined shapes: high, narrow, controlled by local variations in sinuous canyons; wide, dip and strike of the controlling low-gradient tubular passages with bed or bedding plane. The fact el1 iptical or lenticular cross that neighboring passages commonly sections; and vertical shafts up exhibit diverse and seemingly to 60 m deep with almost perfectly independent trends is due to local vertical walls. Canyons and shafts variations in structure from bed are far more common in the to bed. Each bed has its own high-relief Chester Upland than in unique structure imposed by the Pennyroyal Plateau. Even depositional irregularites and by vertical shafts are strongly variations in thickness. influenced by the prominent Generalized contour maps of the bedding , because inf 1owing water geologic structure drawn on a is generally perched along single stratigraphic horizon are bedding-plane partings or on a rarely of use in interpreting the relative1y resistant bed. Growth local structure that controls cave of a typical shaft takes place passages, however useful they mav downward in stages through time, be in regional studies. with the bottom deepening from one Network mazes are rare in the major bed to the next, and with Mammoth Cave area because of the CAVE DEPOSITS

lack of prominent joints and the Calcium carbonate concentraton of groundwater such as f 1owstone and dripstone recharge into numerous small point are most common where abundant but sources (mainly sinkholes). Small diffuse water passes through soil networks do occur where the upper into limestone along the flanks of beds of the Girkin Formation are ridges and knobs in the Chester highly fractured and overlain by Upland. This water picks up a thin permeable sandstone, such as great deal of carbon dioxide from in parts of James Cave. In the soil and quickly approaches general, the insoluble caprock of equi 1i brium with dissolved the Chester Upland is too thick limestone at high concentration. and i mpermeabl e to admi t enough The water degasses readily when it water to form caves. Instead, the enters underlying dry, aerated caprock forms a barrier to all but passages, which have a much lower diffuse capillary water, so carbon dioxide content. In the underlying upper-level passages no Pennyroyal the frequent flooding longer occupied by streams are of passages is not so conducive to exceptionally dry. The thin the growth of speleothems, limestone units interbedded with a1though 1ocal areas of the insoluble Chesterian rocks occur, especi a1 1y in dry upper actually reduce the amount of cave levels. Isotopic analysis of water passing downward to the main carbonate speleothems from the limestone below by shunting water Mammoth Cave System has been 1aterall y to perched springs. useful in interpreting past

temperatures of the region- (Harmon CAVE DEPOSITS and others, 1978). The small amount of water that Detrital sediment is the most penetrates the caprock of the common type of cave deposit in the Chester Upland generally moves by region (see Davies and Chao, 1959; differences in capillary Collier and Flint, 1974). Present potential, rather than by gravity. or former stream passages that is drawn toward dry, aerated have had rapid flow contain sand It caves from the surrounding moist and sandstone fragments mainly and from sandstone in the Chesterian 1i mestone particularly gypsum caprock, quartz pebbles from the and epsomite, are deposted in basal Pennsylvanian rocks (1imi ted these passages (Pohl and White, 1965). For a complete description almost exclusively to caves in the of these materials, Hill Chester Upland), and chert (1976). fragments from the limestone. Much of this material is second-generation sediment, having HYDROLOGIC CONTROLS collected first at the surface at the base of steep slopes and later OF CAVE PATTERNS carried underground through Although all the caves of the sinkholes. Silt and clay from region share certain broad heterogeneous sources collect in similarities, there are some passages that are flooded by striking differences between slow-moving water. These include caves, and commonly between passages abandoned by low-flow different parts of the same cave streams but which are subject to system, that are caused by f 1ooding by slow-movi ng water variations in the hydrologic overflowing from active passages setting. These differences are during high f 1ow, and base-1 eve1 most easily seen in the Mammoth passages backflooded by nearby Cave area, where the great size of rivers. Caves of the Pennyroyal the drainage basins and length of Plateau are particularly rich in the caves have a1lowed the +ullest silt, clay, and residual chert possible response to variations in derived from the local limestone. 1ocal hydrology. THE MAMMOTH CAVE REGION AND PENNYROYAL PLATEAU 111

Some caves, or sections of strong tendency to perch on caves, such as the upper levels of relative1y insoluble beds or along Mammoth Cave {Fig. 11 ) , consist of bedding-plane partings. Vadose a few major passage5 that extend cave passages tend to be dip for great distances with little oriented and may extend several change. Some, such as Crump Spring kilometers before they merge with Cave, are complex tangles of present or former phreatic narrow canyons and rudimentary passages. While a vadose passage tubes. Others, such as Hicks Cave, is forming, it tends to lose water are complex anastomotic systems through its floor along fractures with numerous overflow routes. or partings that eventually These differences can be enlarge enough by solution to attributed mainly to the divert the entire stream flow into individual hydro1ogi c setting. a lower-level route. With time, a single input of vadose water can Caves in High-Level create a system of canyons on many Recharge Areas different levels, each in a Vadose flow is drawn downward different bed or series of beds. by gravity along the steepest Variations in dip from one bed to avai 1able path. ' In the Mammoth another cause each canyon level to Cave area the prominent bedding follow a slightly different course inhibits the direct vertical from its predecessor. descent of water along discordant The higher the input above the fractures, and instead water has a local base level. the greater the

Figure 11. Audubon Avenue in Mammoth Cave is a typical upper-level canyon of late Tertiary or early Quaternary age, partly filled with detrital sediment. CAVES ALONG MAJOR PHREATIC DRAINAGE LINES potenti a1 for mu1 ti-1 evel vadose during lengthy periods of stable canyons to develop. The most base level (Fig. 12i. Those favorable of such inputs are passages that do loop downward located in the Chester Upland into the phreatic zone remain along the flanks of ridges or in active and continue to grow long karst valleys, where 50 to 100 m after contemporary passages formed of re1i ef ex isted above base 1evel at the water table have been while the caves were forming. The abandoned by a lowering of base Mammoth Cave System, for instance, level. The large size of the contains local areas where vadose passages in and around Echo River canyons, shafts, and perched tubes in Mammoth Cave is due partly to are so densely concentrated that the fact that they are they are almost impossible to contemporaries of passages that portray clearly on a plan-vi ew lie as much as 25 m higher. map. Each of these sections has Although the higher passages have been formed by only a few long been dry, the lower parts of concentrated high-level inputs of the passage loops are still active aggressive vadose water where a today and have grown to a stream valley has breached the considerably larger size than insoluble caprock. The complexity otherwise would have been the of the caves is due to the case. tendency for vadose water to Many caves in the Pennyroyal divert to lower routes, and to the Plateau have the re1ativel y simp1 e many thin beds that provide those morphology described here. Because routes, rather than to a o-f the low relief, vadose feeders complicated geomorphic history or reach the water table over rather to shifts in the pattern of short distances, in comparison . with those in the Chester Upland, so complex systems of canyons and Caves Along Major Phreatic shafts are rare. Drainage Lines Caves at the Downstream Ends Phreatic cave passages form zones of low hydraulic head, of Catchment Areas toward which the water in An exception to the simplicity surrounding openings is drawn. As of caves in the Pennyroyal Plateau a result, water in these passages is introduced by flood water has little or no tendency to overflow. In caves subject to divert to new routes. 6s long as flooding, which is somewhat more they are located at or below the common in the Pennyroyal than in local river level, no matter how the Chester Upland, the main large they grow or how much flow passages fill with high-pressure they acquire, their position floodwater much faster than the remains stable. Such passages water table can rise in the usually have a tubular shape and surrounding fractured but extend for -great distances with non-cavernous bedrock. When this little change (Fig. 10). They may happens, the hydraulic gradients be located precisely at base level around the pasages are reversed, and be filled with water on1 y so they are no longer toward the during periods of high flow, or passages but away from them. Water they may descend into the phreatic is forced out of the cave passages zone and be perennially water into fractures in the surrounding filled over part or all of their 1imestone, and since this water is length. The tendency for water to soluti onall y aggressive it tends follow shallow paths near the top to form a1ternate routes of flow. of the phreatic zone, rather than In a rather short time, penetrate to considerable depth, geological 1y speaking, this leads to the development of high-gradi ent aggressive water can distinct levels of major passages form a system of diversion I THE MAMMOTH CAVE REGION AND PENNYROYAL PLATEAU 113 I

I 650 -- Blue Spring Branch Gothic Ave. Broadway

600 -

Cleaveland Ave. - 175

-2 -2 550 -

Gt. Relief Hall

500 - - 150

M.

450 - F T.

Figure 12. Generalized profile through part of Mammoth Cave, showing the three major levels of passage development (labeled 1, 2, and 3). Shaded areas represent the original tubular portions of each passage. Passages on level 3 were originally adjusted to a static base level at 152 m, but the phreatic sections span a vertical range of 20 m. See text for further details. passages emanating from the interconnections between passages original ones, either connecting in the area, dye introduced at a different parts of caves (as in sing1e upstream point not Parker Cave) or forming surprising1y wi11 usual 1y emerge distributaries to the nearest at many different springs, river (as in Hicks Cave). These especially during high flow. passages typically form Further details on the approximately at the same level morphology and geology o# caves in and along the same limestone beds the region are given by White and as the host passages, so the others (197U), Ewers and Quinlan resulting pattern is distinct1y (1981), Palmer (19811, and Quinlan anastomot ic. and others (1983). The processes described here are most common in the downstream RELATIONSHIP OF KARST TO THE parts of large drainage basins, LOCAL GEOMORPHIC HISTORY where the discharge is greatest Two concepts must be integrated and where phreatic passages are when interpreting the geomorphic most numerous. Distributaries evolution of the region. First, in around springs are favored by the the broadest sense, this evolution tendency for blockage of out1ets can be viewed as a steady-state by collapse material and landslide system of gradual uplift and debris. Fackf looding of caves from erosion, in which the mainly the adjacent river may contribute clastic Chesterian and to the enlargement of a Pennsylvanian rocks are gradually distributary system. 61luviation stripped from the limestone, of local river valleys has caused causing the locus of karst many passages in the downstream development to migrate with time ends of cave systems to be toward the northwest in the re-flooded. Although they may downdip direction. With this view, represent several different stages the present karst landscape and of cave origin, all passages below caves are seen to represent (1) a present base level are now water youthful stage in the northwestern filled. Because of the numerous areas where the caprock is just RELATIONSHIP OF KARST TO THE LOCAL GEOMORPHIC HISTORY beginning to be breached by The uppermost levels in Mammoth erosion, (2) an early-mature Cave and surrounding caves formed stage, as in the Mammoth Cave at this time. They reflect this area. where limestones are exposed slow degradation and aggradation with maximum relief and cave in that they are wide, large tubes development is greatest, (3) a and canyons up to 25 m deep, late-mature stage, as in the filled with sediment to at least sinkhole plain, where the caprock two-thirds of their depth in many has been completely removed, places (Fig. 11) . These passages a1 1owing f ull development of are concentrated at altitudes surf ace karst but a diminishing around 182 to 190 m at their vertical extent of caves, and (4) downstream ends near the Green an old-age stage in the River, more or less at grade with southeastern areas where the most nearby parts of the Pennyroyal soluble limestone has been removed surface. They are relatively few, by erosi on. because the limestone was sparsely dissected at that time, and Superimposed upon this general underground drainage was fed only and steady-state trend is the by a few large sinking streams in influence of relatively short-term the karst valleys and adjacent changes in climate and base level. Pennyroyal Plateau. The Pennyroyal The otherwise smooth transition at this time supported mainly from one geomorphic stage to surf ace drainage, as the slow another is periodically arrested erosion of the region promoted low or accelerated, leaving a unique relief with on1 y small local karst signature on the karst features basins (Miotke and fapenberg, that form at any given time. 1972). The landscape was probably Although the steady-state model not very different from that of must be kept in mind as a general the Great Val ley of Pennsylvania backdrop, the secondary aspects of and Virginia on and geomorphic history are especially Ordovician carbonate rocks today. important here. The specific morphology of caves and other There has been some debate as karst features tells a far more to the origin of the flat surface varied and subtle tale than does of the Pennyroyal Plateau. Its the steady-state model, and it is approximate concordance with the more fruitful in the strata and the presence of interpretation of past geomorphic low-re1 ief areas under1ai n by conditions. cherty horizons has led some authors to interpret the During the late Tertiary and Pennyroyal as a stripped early Quaternary periods, the structural plain (e.g., Quinlan, region underwent slow erosional 1970) . Another aspect of geologic degradation, which alternated with control is the strong relationship periods of broad-scal e between sinkhole distribution and aggradation. This a1ternation stratigraphy (Howard, 1968). between erosion and deposition was Others, such as Miotke and probably caused by cyclic changes Papenberg ( 1972) and We1 1s i 1976) , in climate from humid to arid. As point to the subtle discordance a result , a 1ow-re1 ief 1andscape between the surface and the strata was developed on the exposed as evidence for base-level limestone, close to base level. control. Both views have merit. It This landscape was the forerunner seems likely that the Pennyroyal of the Pennyroyal Plateau. The surface formed close to fluvial area containing the insoluble base level during the slow caprock projected as a resistant Tertiary dissection of the region, hilly region, the forerunner of but that local areas show the the Chester Upland, but with far effect of differential resistance less relief than now. of strata. THE MAMMOTH CAVE REGION AND PENNYROYAL PLATEAU

During the Tertiary Period, a few large passages that are most of the surf ace drainage from still forming today, many small the Appalachian Mountains drained ones are being formed by numerous through the Teays River, located local sources of recharge from north of the present Ohio River. karst valleys and ridge flanks. Early in the Pleistocene, this Interpretation of passage drainage was diverted southward levels in Mammoth Cave is not a into the Ohio River by continental simple matter of charting the glaciers. The resulting increase elevation of the largest passage. in discharge caused the Ohio to Most passages in the system have entrench rapidly. Consequently, both an upstream vadose section the Green River and other and a downstream phreatic section tributaries of the Ohio deepened formed at or below the water rapid1y as well (Miotke and table. The elevation at which this Palmer, 1972) . Deep, steep-wal led transition takes place is the most valleys were produced in the reliable indicator of a relative1y Pennyroyal Plateau and Chester static base level. at the time a Up1 and. This entrenchment was given passage was forming. Figure periodically interrupted by 12 shows the major passage levels periods of aggradation, probably in Mammoth Cave and some of the coinciding with periods of complex ities that make their glaciation. In limestone areas, interpretation difficult. The while major streams such as the uppermost level 182 to 190 m is Green River became entrenched, best represented by Gothic Avenue minor tributaries remained and its upstream and downstream hanging, and water diverted extensions. No vadose section is underground. The purer limestone accessible, and the entire passage of the low-relief Pennyroyal consists of an undulant tube. surf ace, especi a1 1y that closest Although it is more or less to the entrenched rivers, concordant with the bedding oi the developed underground drainage, lower Girkin Formation, it rises caves, and sinkholes. Karst and falls imperceptibly along its valleys formed between ridges in length. Leveling surveys show that the Chester Upland. there is virtully no dip on the Rapid shifts in base level beds, except for minor local during the Quaternary caused cave structures. The downstream end passages to form quickly at many (Broadway and Audubon Avenue) is different levels (Figs. 8 and 10). rough1y para1lel to the local Most prominent are those at 168 m strike, but otherwise the passage and at 152 m. These levels shows no systematic relationship probably coincide with periods of to the limestone structure. The re1ati vel y stab1e base 1eve1 , persistent narrow elevation range rather than with favorable rock of the passage and its tubular units, because the passages that shape (except where vadose cluster at uniform elevations entrenchment has taken place in occur at different stratigraphic the downstream sections) suggest horizons. Paleomagneti c anal ysi 5 development very near the water of sediment has shown that the 152 table during a period of nearly m level is at least 700,000 years static fluvial base level. This old (Schmidt, 1982). data is strengthened by the fact With time, greater dissection that other major passages in the and relief caused an increase in system reach their maximum width the number of recharge points. at this same elevaion. These Although major flow routes from passages include the main passages the Pennyroyal Plateau sti11 of Salts Cave and Crystal Cave in exist, recharge from the Chester the northeastern part of Flint Upland has become divided into Ridge. They lie in rock units many small inputs. In addition to different from those of Gothic RELATIONSHIP OF KARST TO THE LOCAL GEOMORPHIC HISTORY

Avenue, so the correspondence of the beds. At the transition point elevation is not caused by representing the former water favorable strata. table, the canyon changes There has been recent gradually to a tube oriented speculation that the uppermost nearly parallel to the local levels of Mammoth Cave were formed strike. This pattern is repeated by paragenesis--this is, as tubes at the same elevation, but in deep in the phreatic zone that entire1y different beds, in were enlarged upward owing to Waterfall Trail and Flint Ridge sediment accumulation on their (Niotke and Palmer, 1972). The floors (Derek C. Ford, Hamilton, upstream end of Waterf a1 1 Trai 1 Ontario, personal communications, still contains an active stream 1975) . Under these conditions a perched on re1ati vel y insoluble tubular-shaped passage would beds, but its former transition migrate upward to the water table, from vadose to phreatic conditions forming a high, narrow canyon is clearly detected by the whose lower parts were filled with transition from dip-oriented sediment. The highest passage in canyon to undulant tube crudely the system, Collins Qvenue in oriented along the strike. Turner Crystal Cave (Flint Ridge), may Avenue in Flint Ridge is another have formed at least partly in example of a tube at this same this way, as it has an irregular level (Fig. 8). ceiling and its accessible part The lowest major level is at rises slightly in the downstream 152 m. Further entrenchment below direction. Paragenesis is unlikely the 167 m level allowed the water in most other passages in the in Boone Avenue to bypass system, however, because of the Cleaveland Avenue and continue on f 01 1owing evidence: ( 1) numerous down the dip as a canyon known as passages are filled entirely to the Pass of El Ghor. At the new the cei 1ing with coarse-grained static water level of 152 m, the sediment capped by a layer of silt canyon gradually changed to a tube and clay only a few centimeters (modified somewhat by later thick, indicating that rapidly entrenchment) in Silliman Avenue. flowing streams deposited the fill Instead of maintaining this level, but did not cause significant however, the tube extended as much upward solution, (2) canyon as 20 m beneath the water table, ceilings are highly concordant to forming the Echo River passage. It the bedding, which is unlikely in rose in the downstream direction a tube dissolved upward through into River Hall, where it joined the phreatic zone, (3) tubular the passage from Great Relief passages on the upper level, such Hall, a major tube at 152 m. The a5 Gothic &venue, are a1so f i11 ed combined passage unfortunately with stream-deposited sediment in terminates in breakdown at River places that have not yet been Hall, so the downstream exhumed by later entrenchment, and cont inuat ion is obscured. However, (4) cut-and-fill structures in the the presence of the Great Re1 ief sediment of upper levels in Long tube at 152 m and the transition Cave, south of Mammoth Cave, show from canyon to tube in Sil.liman evidence of meandering channels Avenue at the same elevation is formed by free-surface streams compelling evidence for a static (James Currens, Kentucky base level at that elevation. Echo Geological Survey, personal River, the low point in this communicaton, 1983). passage system, has therefore been The level at 167 to 170 m is actively forming ever since the best observed in Cleavel and Green River was at 152 m. The Avenue. Boone Avenue, its upstream large size of the Echo River end, is a vadpse canyon that passage can be accounted for in extends directly down the dip of that way. Other passages at the THE MAMMOTH CAVE REGION AND PENNYROYAL PLATEAU

152 m level include Floyd's Lost George, A. I.,1976, Karst and Passage and Swinnerton Avenue in cave distribution in north- Flint Ridge. Each show a transiion central Kentucky: National Spe- from vadose to shallow-phreatic at leological Society Bulletin, v. the same elevation but in 38, p. 93-98. different rock units. The great extent of caves and Harmon, R. S., Schwarcz, H. P., karst features in the Mammoth Cave and Ford, D. C., 1978, Stable area is clearly not a simple isotope geochemi stry of spel e- product of the vast exposure of othems and cave waters from the limestone, and the steady-state Flint Ridge-Mammoth Cave System, model of slow denudation is Kentucky: Implication for ter- inadequate to explain it. Without restrial climate change during the relatively rapid and deep the period 230,000-100,000 years entrenchment resulting from BP: Journal of Geology, v. 86, Quaternary changes in the Ohio p. 373-384. River drainage, the region would probably contain only a small Hess, 3. W., 1976, A review of the fraction of the ,caves and karst hydrology of the Central Ken- topography that it does now. tucky Karst: National Speleolog- ical Society Bulletin, v. 38, p. REFERENCES 99-102. Borden, James, and Crecelius, Hill, C. A., 1976, Cave minerals: Peter, 1984, The Roppel- Austin, Texas, Speleo Press, 137 Mammoth Connection: National P Spel eologi cal Soci ety News, v. - 42, p. 103-109. Howard, A. D., 1968, Stratigraphic and structural controls on land- Brucker, R. W. and Watson, R. A., form development in the Central 1976, The longest cave: New Kentucky Karst: National Spele- York, Alfred A Knopf, 316 p. ological Society Bulletin, v. 30, p. 95-114. Collier, C. R., and Flint, R. F., 1974, Fluvial sedimentation in Lawrence, J., Jr., and Erucker, R. Mammoth Cave, Kentucky: U. S. W., 1955, The caves beyond (1975 Geological Survey Professional reprint): Teaneck, New Jersey, Paper 475-D, p. 141-143. Zephyrus Press, 283 p.

Davies, W. E., and Chao, E. C. T., Miotke, F. D. , and Palmer, A. N. , 1959, Report on sediments in 1972, Genetic re1ationship Mammoth Cave, Kentucky: U. S. between caves and landforms in Geological Survey Admi nistra- the Mammoth Cave National Park tion Report, 117 p. area: Wurtzburg, Germany, Eohler Verlag, 69 p. Deike, G. H., 1967, The develop- ment of caverns of the Mammoth Miotke, F. D., and Papenberg, H., Cave Region: University Park, 1972, Geomorphology and hydrol- Pennsylvania State University, ogy of the sinkhole plain and Ph.D. Dissertation, 235 p. Glasgow Upland, Central Ken- tucky Karst, preliminary report: Ewers, R. 0. , and Quinlan, J. F. , Caves and Karst, v. 14, p. 25- 1981, Cavern porosity develop- 32. ment in limestone, A low dip model from Mammoth Cave, Ken- Palmer, A. N., 1977, Influence of tucky: Eighth International geologic structure on ground- Congress Spel eol ogy , Bowl ing water flow and cave development Green, Kentucky, v. 2, p. 727- in Mammoth Cave National Park, 731. Kentucky, U. S. A. : International REFERENCES CITED

Association of Hydrogeol ogi sts, Karst: Phase I: University of 12th Memoirs, p. 405-414. Kentucky, Water Resources Re- search Institute, Research Palmer, A. N., 1981, A geologic Report 101, 93 p. guide to Mammoth Cave National Park: Teaneck, New Jersey, Buinlan, J. F., and Ray. J. A., Zephyrus Press, 196 p. 1981, Groundwater basins in the Mammoth Cave Reqion, Kentucky: Pohl, E. R., 1970, Upper Missis- Fri ends of the Karst, Occasional sippian deposits in south- Publication 1. central Kentucky: Kentucky Academy of Sciences Trans- Schmidt , V. A. , 1982, Magnetostra- actions, v. 31, p. 1-15. tigraphy of sediments in Mammoth Cave, Kentucky: Science, v. 217, Pohl , E. R., and White, W. B., p. 827-829. 1965, Sulfate minerals: their origin in the Central Kentucky Swann D. H. 1964, Late Missi 5- Karst: American Mineralogist, v. , , sippian rhythmic sediments of 50, p.1462-1465. the Mississippi Val ley: Associa- tion of American Petroleum Geol- Quinlan, J. F., 1970, Central Ken- ogists Bulletin, v.48, p. 637- tucky Karst: Reunion internat- 658. ional e karst01ogie en Languedoc- Provence 1968, Actes, Mediter- Watson, R. A. 1966, Central Ken- anee, studes et traveaux, v. 7, , tucky Karst hydrology: National p. 235-253. Speleological Society Ful letin, v. 28, p. 159-166. Ruinlan, J. F., Ewers, R. O., Ray, J. A., Powell, R. L., and Krothe, N. C., 1983, Ground- Wells, S. G., 1976, Sinkhole Plain water hydro1ogy and geomorphol- evolution in the Central Ken- ogy of the Mammoth Cave Reqion, tucky Karst: National Spel eol og- Kentucky, and of the Mitchell ical Society Bulletin, v. 38, p. Plain, Indiana: Geological Soci - 103- 106. ety of America Annual Meeting, Guidebook to Field Trip 7. 85 p. White, W. B., Watson, R. A., Pohl, E. R., and Brucker, A. W., 1970, Ouinlan, J. F. , and Rowe, D. R., The Central Kentucky Karst: 1977, Hydrology and water qual- Geographical Review, v. 60, p. ity in the Central Kentucky 88-1 15. Chapter 8 WESTERN KENTUCKY REGION John Mylroie Department of Geosciences Murray State University Murray, Kentucky 42071

and

Mike Dyas National Speleological Society 6009 Backlick Road Springfield, Virginia 22150

The Western Kentucky Region is and in the northwest by tribu- a major cave area of the United taries of the Ohio River; but the States, but only recently has its majoriy of the area's drainage is potential been appreciated. The into the Cumberland River by area has been somewhat lost tributaries of the Red River and between more glamorous areas the Little River. around it, such as Perry County, The western Kentucky area has , to the west; Monroe over 200 described and documented County, Illinois, to the north: caves, with 147 currently surveyed the Mammoth Cave Region of Ken- to an aggregate length of almost tucky to the east; and the 100 km. Lisanby Cave, Caldwell Cumberland Plateau area of Tennessee to the southeast. THE WESTERN KENTUCKY REGION The Western Kentucky Hegion can be defined as the Mississippian limestone area of Kentucky border- ing the Western Kentucky Coal Field and west of the West Fork of Drakes Creek, a tributary of ')..I\ I WESTERR KENTUCKY COAL FIELD \ the Barren River; it is part of the Interior Lowlands Province of the ea.stern United States (Fig. 1). Going east to west, this area includes parts of Warren and Simpson counties, minor areas of Butler and Muhlenburg counties, and large areas of Logan, Todd, v-.-. ?-.: i j I.-.+- 1.- Christian, Trigg, Caldwell, Lyon, '. ! TEMNESSEE Crittenden , and Livi ngston -.- counties. The western Kentucky Figure la. The Western Kentucky area is drained in the northeast by tributaries of the Green River Region. 120 WESTERN KENTUCKY REGION

MAJOR STREAMS AND CAVES OF THE WESTERN KENTUCKY REGION

LOCATION OF CAVES DISCUSSED IN TEXT MAJOR AREA STREAMS 1 - & ANGEL CAVES A - WEST FORK DRAKE'S CREEK 2 - GORHAM CAVE B - BARREN RIVER 3 OAKVILLE CAVE kilareters - C - GREEN RIVER 4 - ANTIOCH CREEK CAVE D - MU0 RIVER 0 25 50 5 - LOVELL CAVE E - RED RIVER 6 - GLOVER'S & TWIN LEVEL CAVES F - ELK FORK, RED RIVER 0- 7 - BIG SULPHUR CAVE G - BIG WEST FORK, RED RIVER 8 - COOL SPRING, DECIBEL a TWIN CAVES H - POND RIVER 9 - HUSK CAVE I - LITTLE RIVER 10 - HARMONY CHURCH CAVE SYSTEM J - SINKING FORK, LITTLE RIVER 11 - SKINFRAME SINKS CAVE K - MUDDY FORK. LITTLE RIVER 12 - MILL BLUFF CAVE L - LIVINGSTON CREEK 13 - LISANBY CAVE M - TRADEWATER RIVER 14 - BAKER CAVE N - CUMBERLAND RIVER 0 - LAKE BARKLEY

..,I. ,..,.. -..-

Figure lb. Major streams and caves of the Western Kentucky ~egion.

County, at 11.3 km, is the longest University at the western edge of cave in the region, and other the region. The students explored caves making the 3.0 km limit for many caves in the area and pro- inclusion in the International duced map5 of a few of the major Long Caves List are shown in caves, such as Cool Spring Cave Table 1. and Glover's Cave. The grotto did The known described and sur- not survive the graduation of its veyed caves are believed to be principal leaders and was dis- only a fraction of the caves in persed bef ore 1975. Local resi- existence in the area. Organized dents made many discoveries and caving came to the Western Ken- explorati on5 during the same time tucky Coal Field Region first in period, most notably Mark Caldwell the late 1960s as the Southwest and colleagues out of Paducah, and Kentucky Student Grotto (SWKSG) , Herb Scott and others out of which operated out of Murray State Hopkinsville. Some survey and WESTERN KENTUCKY REGION

Table I Murray State University, and was recruited into the fledging WKSS. INTERNATIONAL LONO* CAVES To avoid the problem of lost data IN WESTERN KENTUCKY and duplication of effort, it was CAVE LENGTH (KM) COUNTY decided to publish an annual Llsanby 11.31 Caldwell report containing descriptions, surveys, and scientific articles Cool Sprlng 5.30 Trl99 on the work done in the WKSS area. Twin Level 5.00 Chrlstlan and Todd Reports covering each year of Blg Sulphur 4.80 Trlgg activity have been published, with Savage 4.28 Logan a limited and controlled

Glover's 3.35 Chrlstlan and Todd distribution to protect the caves. A compilation of the WKSS activity Gorham 3.21 Logan through January 1985 is shown in Sklnframe Slnks (Rlce) 3.00 Caldwel l Table 2.

*Caves 3 km or more in length. GEOLOGY Western Kentucky may be viewed as a series of cuestas formed by documentation were done by these the removal of strata on the informal groups, but the Western southern and western flanks of the Kentucky Coal Field Region was Western Kentucky Coal Field. The basically a blank spot within the resulting escarpments define spel eologi cal know1edge of the up1ands and low1ands of distinct United States. Some of the best geologic and geographic character. data were in the hands of speleo- The western Kentucky karst is a biologists, such as Tom Earr and sinkhole pldin developed on the John Holsinger, who made forays into the area for specimen collec- tion, and the Evansville Metropolitan Grotto, which did Table I1 extensive work in the Glover's Cave area of Chr isti an County. WKSS ACTIVITY THROUQH 1985 Significant progress in the study and documentation of the COUNTY NUMBER OF NUMBER KILOMETERS caves and karst of western CAVES KNOWN MAPPED MAPPED Kentucky began in August of 1977 with the establishment, by the Butler 2 0 0 Board of Governors of the National Caldwell 36 22 23.42 Speleological Society (NSS) of the Christian 20 Western Kentucky Speleological 12 13.24 Survey (WKSS). Mike Dyas, a caver Crittenden 14 9 2.93 with extensive experience in other Livingston 14 7 2.19 regions, began caving in western Kentucky in 1975 when his family Logan 52 3 1 19.53 bought land in Caldwell County. Lyon 9 1 0.6 Mike's interest in the region caused him to search for available Muhlenberg 2 2 1.13 data from the NSS and elsewhere. Simpson* 1 1 0.01 It became immedi ate1y obvi 01-15 that Todd there was little published 16 11 6.28 information, yet field work showed Trlgg 5 1 50 24.09 a region with considerable Warren* 1 1 0.33 potential. In the spring of 1977. John Mylroie, a longtime caver Total 218 147 93.75 from the Northeast, accepted a position in the geology program at *Portion of the county within the WKSS only. 122 ST. LOUIS LIMESTONE

Pennyroyal Plateau, a long band of the faults general 1y strike Mississippi an 1imestones, northeast and are related to extending from eastern Kentucky, similar faults in southern arching southward through central Illinois. The Illinois-Kentucky and western Kentucky and Fluorspar District has produced neighboring Tennessee, to southern fluorite and sphalerite from Illinois and eastern Missouri. In mineral ization associated with western Kentucky, the limestones these faults. Faults in Caldwell dip approximately 0.5" to the and Christian counties have a north-northeast but are locally northwestly or more westerly undulatory. The plateau is gently trend. To date no significant rolling to nearly level, averaging fluorite mining has taken place no more than 30 to 40 m in relief, along these faults. The gentle with average elevation around 150 regional dip for the rocks in this m. While the plateau length is area changes from northly in the extensive, its width varies southern portion to northeastly as considerably. In western Kentucky, one moves northwestly around the the plateau averages 30 km in western flank of the Illinois width. Commonly referred to as the Fasin and the encl osed Western "Sinks" or the "Sinkhole Plain," Kentucky Coal Field. the plateau is characterized by For the most part the rocks in thousands of shallow sinkholes, this area are of Mississippian age occasional blind valleys, and and belong to the 'Meramecian and numerous karst windows. Karst Chesterian stages. In extreme pavement has been found in only a western Trigg County, Cretaceous few locations, due to the deposits overlie the Mississippian extensive soi1 development in the rocks. Along the shores of region. Stream and river valleys Kentucky Lake and Lake Barkley, are we1 1 a1 luviated with sands, 1imestones of older age are clays, and chert gravel. Sticky present locally. However, the vast red clay and chert cobbles mantle majority of the known caves are the ridge tops and plains. found in or above the St. Louis Internal drainage is well Limestone (Fig. 2) . Lithologic developed with surface drainage descriptions and correlation being restricted to several information can be found in the entrenched base 1eve1 streams. 7.5 minute geologic quadrangle The 1imestones of western maps of this area. Detailed Kentucky are a1 l Mississippian chemical analyses and 1i thologic (Carboniferous) , and are simi lar descriptions are also available in in many characteristics to other Dever and McGrain (1969). Figure 2 Missi ssippi an 1i mestones of and the accompanying lithologic Kentucky and surrounding states. descriptions are a summary of that The following geologic description information. is condensed from Whaley and Black !1978). St. Louis Limestone The age and general structure The St. Louis Limestone is a of the rocks in Western Kentucky light-brown to gray, fine to can be seen on the Geological Map coarse-grai ned , argi 11 aceous of Kentucky (McFarlan and Jones, limestone, 120 to 150 m thick. 1954). Dolomitic 1imestone and beds of Faults in the southern portion oolitic limestone are present. of the area (Logan, Todd, Trigg, Beds are 2 to 60 cm thick. The and southern Christian counties) unit is divided into upper and are less numerous than faults in 1ower 1i mestone members. The upper northwestern portion (Christian, member contains dark gray chert Lyon, Cal dwell , Livi ngston , and nodules and in part of the area it Crittenden counties). In has been mapped as the lower unit Livingston and Crittenden counties of he Ste. Genevieve Limestone. WESTERN KENTUCKY REGION 123

STRATIGRAPHIC SECT1 ON OF WESTERN KENTUCKY

GOLCONDA FORMATION

---_ ----_ SANDSTONE

PAINT CREEK

FORMATION _------_

STE. GENEVIEVE

STL GENEVIFNE

LOGAN --- -_ COUNTY COUNTY

CALDWELL COUNTY CHRISTIAN COUNTY TODD COUNTY

Figure 2. Stratigraphic section of western Kentucky.

Ste. Genevieve Limestone 1 i mestone with grayi sh-green shale The Ste. Genevieve Limestone is laminae. Chert nodules occur in divided into three members: the the lower beds. lower 60+ m thick Fredonia Limestone, the middle 3+ m-thick Renault Limestone Hosiclare Sandstone, and the upper The Renault Limestone is a 8+ m-thick Levias Limestone. The light to medium-gray, variable light gray Fredonia argillaceous, oolitic, fine to Limestone is predominant1 y an medium-grained 1 imestone, that is oolitic limestone with beds of local 1 y i nterbedded with shale and fine crystalline, dolomitic si 1 tstone. The formation averages limestone and medium to coarse 20 m in thicknrss but thins to 10 crystalline, fossiliferous m in the central portion of the limestone. The Hosiclare Sandstone region. In the eastern portion of grades into shale in the southern the region the Renault Limestone . part of the region. The Levias is equivalent to the lower 15 to Limestone is a light gray, oolitic 18 m of the Girkin Formation. BETHEL SANDSTONE

Bethel Sandstone and shales grade eastward into the Member. The The Bethel is light-brownish to Beech Creek is a thin, fine to gray, fine to medium-grained, thin coarse-grained, fossiliferous to thick-bedded sandstone. 1imestone. The Haney is an Siltstone and shale are found in argi1 laceous 1imestone that yields the upper portion of this unit. residual chert upon . The Bethel exceeds m in 35 The Big Clifty is a fine to very thickness in the west, but fine-grained, thin to very gradually thins to the east where thick-bedded sandstone. Thin beds it grades into the Girkin tend to be ripple marked while Formati on. thick beds contain cross beds. Paint Creek Limestone Siltstone and shale are interbedded in the sandstone and The Paint Creek is a variable locally may constitute the major unit consisting of limestone, lithology of this member. shale, and sandstone. Fine to medi um-grai ned, f ossi 1if prous, The placement of the major oolitic, and argillaceous beds rivers has resulted in only a.few make up the limestone components valleys that cut through both the that dominate in the east. Chert overlying clastic rocks and deep lenses are locally present. Shale into the limestones, as the Green and sandstone are dominant in the River does at Mammoth Cave. western part of the area. This Therefore, cave development has unit is part of the Girkin progressed poorly in the Formation in the eastern part of limestones under the the region. sandstone-capped up1apds, with small maze caves and short, simple Cypress Sandstone stream caves being the dominant The Cypress Sandstone consists type. On the adjacent sinkhole of sandstone and interbedded shale plain, large, dendritic, base and siltstone. The sandstone is level cave systems are dound. The white to light gray in the west, relief of 30 to 40 m limits the but light-tan to brown in the amount of abandoned upper 1eve1 s east. Ripple marks in the above the active cave system, thin-bedded deposits and although in many caves, two, cross-beds in the thicker beds are three, or four dry, abandoned common throughout the study area. 1evels can be found compressed This unit thickens westward from within 10 m of base level. Caldwell County and temporarily Occasional isolated fragments of thickens eastward where it pinches upper level passages can be found into the Girkin Formation. on the interfluves between .major Thickness ranges from 1 to 35+ m streams, 'qdicating a in the west. well-developed subsurface flow be-fore the incision of the current master streams to their present The Go1 conda Formation consists elevation. In the extreme of limestone, sandstone, and shale northwestern part of the region, units which in the eastern portion the limestone has been broken up of the study area have been by complex normal faulting and assigned member status. Its fluorite mineralization. Cave thickness varies from 27 to 45 m. systems are smaller but often The upper portion of the Golconda, complex and highly influenced by a limestone unit, to the east is structure. equivalent to the Haney Limestone The limestone is covered by a Member while the lower portion, thick residual soi 1. figricultural also a limestone, is correlative land use has mobi 1i zed this soi 1, with the and a1 luviation of sinkholes and Member. The intervening sandstones cave passages is common. Open WESTERN KENTUCKY REGION

Kentucky area, and can infuence cave development on a local scale (Fig. 5). The regional joint pattern is a conjugate set approximately northeast to southwest and northwest to southeast in orientation. The caves of the sinkhole plain are elliptical tubes up to 5 m high an 15 m wide, and the abandoned upper levels, where found, are similar. Tributaries commonly occur as vadose canyons, and in the low gradient of the region they can meander dramatically, producing numerous abandoned sections and areas of substantial complexity (Fig. 6). The canyons are rarely over 10 m high and 3 m wide, and generally break up into impassable inlets at the upstream end. Cross-links between adjacent cave systems occur, such as at the Lisanby-Farless connection Figure 4. Karst features of the (Caldwell County, Figure 7),or at Sinking Fork area, Trigg County, Himstone Runway Cave (Trigg Kentucky. County), not an unexpected occurrence in a low gradient situation where a minor rise in water level can merge subsurface vertical shafts are relative1y drainage basins. rare, and entrances to the caves Along the meanders of the are of three main types: low master streams, piracies or gradient entrances at the stream meander cutoffs are common. They sink points, large collapse sinks have been described from the and karst windows, and active Sinking Fork of the Little River (Fig. 3) or abandoned resurgences (Moore and Mylroie, 1979: see Fig. in the bluffs along the master 4) and the West Fork of the Red streams of the area. , River (Mason. et a1 . . 1984) . Figure 8 shows the relationship HYDROLOGY between meanders, meander cutoffs, Other than data collected by and sinkhole plain caves along the direct exploration, little is West Fork of the Red River. A known about the hydrology of the compl ex re1ationship between river area. Few dye traces have been flow and cave development can run, but where they have, occur. For example, the efficient subsurface flow paths of up to 5.5 and complete flow of surface water km have been delineated (Moore and through a meander neck can have Hylroie, 1979), as in the Sinking major consequences for the karst Fork basin of Trigg County (Fig. features associated with the 4). The regional dip, being low meander loop that has been and undulatory (except in the abandoned. northwest), produces cave system The number of caves located, orientations controlled by the explored, and surveyed is be1i eved location of the master sur+ace to be only a fraction of the streams. Occasional normal faults accesible total for the region are found throughout the western (Table 2). With such a small data 1 126 HYDROLOGY

I

KARST FEATURES OF THE SINKING FORK AREA

TRIGG COUNTY, KENTUCKY

A HILLTOP CAVE 1 SINKING FORK B CUTOFF CAVES 1 STILLHOUSE BRANCH C COOL SPRING CAVE 1 STEELE BRANCH D RIVER ROAD CAVE 4 BOYD LAKE BRANCH E TWIN TUNNELS CAVE F DITCH PIT C DECIBEL CAVE 4% CONTOUR LINE H MILL STREAM SPRING WITH ELEVATION I BOATWRIGHT HOLE N J SKYCOLUMN CAVE 0SURFACE STREAM 1 I( PIPELINE CAVE L BUZZARD CAVE CAVE PASSAGE M INSURGENCE COMPLEX lea 4400 N UPSTREAM INSURGENCE OF + - + - PROVEN SUBSURFACE meters SINKING FORK FLOW ROUTE 0 1200 CONTOUR ELEVATIONS IN FEET

Figure 4. Karst features of the Sinking Fork area, Trigg County, Kentucky. WESTERN KENTUCKY REGION 127

TODD COUNTY. KENTUCKY mwman sPFurucIUL mEN N3mEm 1962 k NxnET, 1963

------U - - ,-FAULT . . RESURGEN~- ,- - EN~ME FIGURE 16

--- CAlWELL COUNTY. KENlUCKl /c- mSlERl m smLnlm1m mw i / ------o-- 1975 - 1983 0 I, qo -- FAULT U I mn- IS Figure 5. Antioch Creek Cave.

base, any generalizations drawn are going to be necessarily imprecise and vague. The principal significance of the region remains the amount of speleology yet to be done.

FARLESS ENTRANCE . -

Figure 7. Lisanby Cave.

WESTERN TODD COUNTY

LEGEND A MIIRI'IIY'S SPRIffi B BUZZARD'S cnvr C CCOAR BLUFF CHURCH CAVE D DRY CIVI E GAIC'S cnvr F DRY FORD CAVE G GLOVLR'S CAVE H RICK'S RISE I WIN LEVFL rAvr J NRNER'S BLUE IIOLI - SIJRiACE SIRLAM /cgiyY i.. CAVE FAS5AGt/ UMlERGRWMl SIRLAM I

Figure 6. Waterfall Canyon area of Lisanby Cave, Caldwell County, Kentucky. Figure 8. Glover's Cave area. MAJOR CAVES OF WESTERN KENTUCKY

MAJOR CAVES OF WESTERN KENTUCKY

Tables 1 and 2 give an idea of LOGAN COUNTY, KY P. 6 S. FORSYIHf M. DYAS the sire and distribution of caves JUNE. NOV 1982 . NOV 1983 in western Kentucky. A complete and detailed description of all 200+ caves of the region is not possible, but abbreviated descriptions of the major caves and types of caves can give an impression of cave development over the entire area. The cave descriptions that follow will 100 progress westward from the Logan I County area to the Cumberland METERS River. Specific location information is deleted in order to preserve the caves. The caves to be described include the longest caves in the region, as seen in Table 1, and other caves that show significant features. The descriptions of the Savage/Angel Cave System, Gorham Cave, Glover ' s Cave, Twin Level Cave, Cool Spring Cave, Big Figure 9, Oakville Cave. Sulphur Cave, Husk Cave, Harmony Church Cave system, Skinframe Sinks Cave, Mi11 Bluff Cave, and stream (same elevation loss in a Lisanby Cave all detail the major shorter linear distance). It has known and active sinkhole plain become f ossi 1 because the section caves of the region. of surf ace stream it delivered The other caves show special water to has become an abandoned features. Oakville Cave (Fig. 9) meander loop in response to a is a phreatic maze with an almost meander cutoff cave (the Cutoff anastomotic nature, further Caves) (Fig. 4) downstream. This complicated by sediment infill and cave shows the complexities of vadose incision. Deci be1 Cave meanders and the interactions they (Fig. 4) is a classic backflood cause when they are pirated maze, located in a hill between a through their meander necks. major stream and its principal Baker Cave, a1though current1y tributary. The cave shows control unmapped, follows a major fluorite and a progressive piracy of the vein for most of its length. It tributary surface stream. has been worked for some of its Antioch Creek Cave (Fig. 5) is fluorite and contains deposits a cave closely controlled by reminiscent of the Derbyshire structure, coincident with or karst area of Great Britain. parallel to a set of east-west Many caves, such as Eaker Cave, normal faults over its 1.5 km Oakville Cave, Glover's Cave, and length, from its upstream sink to others not described in this its downstream resurgence. chapter, are major archeological Twin Tunnels Cave (Fig. 4) 1s a or paleontological sites. While cave demonstrating some important much plundering has gone on, some aspects of stream incision and sites, such as Savage Cave, are meandering in limestones. The cave currently protected and undergoing is a f ossi 1 meander cutoff , formed intensive study. because its route offered a Some of the region's caves have steeper gradient to the surface been major recreational caves, WESTERN KENTUCKY REGION like Glover's, Cool Spring, and From Savage Cave to the suspected Mill bluff Cave, while Love11 Cave resurgence is a distance of over 1 was once commercialized. Cave km. management is in its infancy in The Savage Entrance to Savage this region but will become a Cave is a large, 25 m wide, 4 m greater concern in the future. high opening in a broad, gentle The Western Kentucky sloping sinkhole. The entrance has Speleological Survey is the major an apparent collapse origin, and source of information for caves of formed at a meeting point of the the region. As Table 2 shows, 218 large upper and middle levels of caves are known from the region, the cave. Down the entrance talus, and 210 have a written a very large (15 m wide by 6 m description, with 147 mapped as of highj passage leads off (west) , January, 1985. The respistory for and is the middle level; just this information is the Western above it to the left, another Kentucky Speleological Survey large passage, the upper level (10 Annual Report Series, pub1 ished m by 6 m) , parallels the middle since 1978. Numerous articles on level. The left passage, or upper archeology, paleontology, level, continues for 80 m. , karst hydrology, The middle level passage, or karst geomorphology, and the main passage, heads west off geochemistry have been published the entrance room, just to the in the Annual Report Series. They right and below the upper level. are on file with the U. S. Initially quite large (15 m by 6 Geological Survey, the National m), after 30 m it closes in, Spel eologi cal Society and numerous averaging 10 m by 3 m. The passage cave groups, or they can be is flat floored and easily obtained from the authors. traversable, but shows evidence of complete back+looding. It Savage Cave System continues in this manner for 350 m westward past breakdown, pools, The Savage Cave System, in and to a complete choke. southern Logan County, consists of To the east at the entrance. a over 5 km of the active and short crawl and stoop over abandoned underground stream breakdown leads to the majority of course of Woolsey Creek and its the cave and the Eidson Entrance. tributaries. Figure 10 shows the Over 3.5 km of most1y easy passage. location map of the major features leads eastward. The stream sumps . of the area, and the cave system in this area, and is not seen components. The features of the again until the resurgence. The Savage Cave System are: Savage main route is the middle level, Cave, Angel Cave, Barnes Cave, usually 3 to 6 m wide and 1 to 3 m Woolsey Creek Cave, the sink point high, with the stream either in of Woolsey Creek, and the the floor or meandering in its own resurgence on the banks of the spacious crawl beneath the middle South Fork of the Red River (Fig. level. At the upstream end of the 10). Portions of Savage Cave and cave, a tributary canyon leads to Angel Cave, as well as all of the Eidson entrance, and the main Parnes Cave, are abandoned upper stream sumps shortly thereafter. A level conduits, indicating a long canyon complex exists in this and varied history for the cave area, and the downstream sump of system. Wool sey Creek Cave Angel Cave is less than 300 m away (located just south of the Woolsey to the east. Creek sink point shown in Fig. 10) is a short tributary cave, Angel Cave releasing its water into the Woolsey Creek blind valley. Angel The 9 m wide, 2 m high entrance Cave and Savage Cave are separated is in a sink midway between the by about 300 m of sumped passage. Eidson Entrance to Savage Cave and ANGEL CAVE

SAVAGE CAVE AREA LOGAN COUNTY, KENTUCKY

wEsImN KENmaY SJELDxmICAL SURVEY 1979 - 1980

Barnes Care

0 .5 1 ----a Lilometrrs mEnucmv TENNESSEE

1

Figure 10. Savage Cave area. the sink point of Woolsey Creek. west, the stream emerges from Figure 10 shows the cave location breakdown: this is the underground and trend. At the left corner of course of Woolsey Creek, the same the cave mouth is a dry side large stream seen in parts of passage that goes for 45 m to the Savage Creek. The passage southwest to a breakdown choke. continues 150 m, partly on two The entrance slope opens to a levels, emerging into one end of a breakdown littered, cobble breakdown room in excess of 90 m floored, west-southwest trending long, 6 to ? m high, and not less trunk, initially 9 to 15 m wide than 18 m wide--at one point over and 2 to 3 m high. The main cave 45 m wide. This impressive stream is not seen at the entrance chamber, comprised of large under normal conditions, being breakdown on a gravel base on the under the breakdown. On the right, right and a sprawling silt bank on 36 m from the entrance, is the the left, is the largest known rather inconspicuous connection to room in any western Kentucky cave. the upstream portion of the cave. It is succeeded by another sizable Forty-five meters beyond to the breakdown room, 45 m long. 23 m WESTERN KENTUCKY REGION wide, and 3 to 4 m high. At the 2-m-high side passage with two end of the second major room, the beautiful pools extends west for passage, with ponded water, 200 m. continues 90 m to the terminal From the entrance, the upstream sump. section extends to most of the The lead to the upstream cave. It trends southwest, portion of the cave is a clean carrying the stream, which flood route, soon reaching the averages 10-m-wide by 6-m-high, main stream. One can continue in a with large breakdown blocks and passage varying from crouchway f lowstone at intervals. One sewer to spacious breakdown trunk kilometer from the entrance, the to wet crawlway. This extension stream forks. The right branch skirts a broad, flat-bottomed sink continues west, past more large and is heavily fractured. At the boulders, for another 300 m: the end, 600 m from the entrance, is a cei 1ing height gradual 1y lowers to back entrance and another 300 m of a periodic sump, pushable another passage nearing the sinks of 7.~u0 - m to a sinkhole entrance, and Wool sey Creek. a final sump 140 m further. There The Savage Cave entrance area, appears to be one or two both surf ace and subsurf ace, is an additional segments of cave in a extreme1y important archeological karst window ad joining this point. site (Carstens, 1980). The At the mainstream fork, the Archeological Conservancy worked left branch goes southwest for 800 with the WKSS, the landowner, and m, intersecting a number of side Murray State University to passages of appreciable extent, to preserve the cave. The property a 10 m wide, 60 m long room with has been transferred to Murray several large columns. Just before State University, and access to its end the stream emerges from a the cave is strictly controlled. 40-m-duck-walk to terminal Archeological study of the cave is breakdown. This breakdown plots to continuing . be within a few meters of a logjam in a nearby karst window. Gorham Cave This cave is named after J. H. Oakville Cave Gorham, whose farm included the The location is a deep, brushy entrance. Rlthough local1 y we1 1 sink surrounded by cultivated known, the 3.2 km cave has fields south-southwest of suffered surprising1y little Oakville, Logan County. The vandalism. Gorham Cave is situated entrance, which takes runoff from in east-central Logan County, heavy precipitation, is 3 m wide about 8 km from Russellville. The and 1 m high, and the cave entrance, a 3 m climb-down, is in initially trends west-northwest. a small sink at the contact of the The cave swings south (see Fig. 9) lower member of the Girkin and becomes a very confusing maze Formation and the Ste. Genevieve of small vadose passages, phreatic Limestone (Fig. 11). The entrance tubes, and abandoned trunk room, littered by massive fragments. An active stream is breakdown, intersects a trunk seen in places, only heard in carrying a stream. The downstream other locations. hir flow is noted section trends northwest, in many places. The largest trunk averaging 10 m in width and 3 to 8 fragments are up to 5 m wide and 3 m in height, for 180 m to a sump. m high. There is abundant mud, and This section reportedly connects several bone sites. at a with a resurgence cave at Spring few promising places could easily Acres, a recreation park a short extend the cave. distance west-northwest of the A total of 1.66 km has been Gorham entrance. Thirty meters surveyed in Oakville Cave, with a before the sump, a 6-m-wide by small amount of unmapped passage. OAKVILLE CAVE

GORHAM CAVE AND VlClNlTY

LOGAN COUNTY, KENTUCKY

WESIEIFJKENNaKY~ICALSURYEP 1979

and Cave

meters

Figure 11. Gorham Cave and vicinity.

Oakvil le Cave is believed to be the master drainage from the part of an extensi ve underground Oakvi lle neighborhood. This is drainage net. The big stream likely the same stream that passes briefly seen at the northwestern through a series @f karst windows "end" of the cave is apparently at Spring Val ley Church, some 1.5 WESTERN KENTUCKY REGION km southwest. The ultimate The west-trending trunk is resurgence is another ki1 ometer initial1y of nice proportions--3.5 southwest, at the former site of to 4.5 m wide and 2.5 to 3.5 m the Pleasant Grove Church. high--wi th broad ledges on one or By western Kentucky standards, the other side, the stream flowing Oakville Cave is regarded as dry shal low1y over gravel, and a and unlikely to flood. However, number of speleothems. The passage large quantities of fresh mud were gradually lowers, however, encountered in the cave in May becoming a cobble-strewn crawl way 1983, after a period of unusually with ponded water about 275 m from heavy rain, so it appears that the entrance; passing several most of the cave can flood significant leads, it continues infrequently. west to link with the upstream A large amount of flint (chert) entrance. The upstream entrance, material is evident in the field like the resurgence, is quite adjoining the entrance sink. Local spacious--9.5 m wide and 3 m high. residents have collected many The main passage assumes the arrowheads and other Indian near1y straight east-west beari ng artifacts from this site, which of the downstream cave, starting must have been a large camp. off as a walking trunk for some Several Indian points have been 200 m, then becoming a low crawl found in parts of the cave itself connecting to the rest of the far from the present entrance. cave. Antioch Creek Cave Lovell Cave The location of this cave is Located south of Weir in roughly 4 km northeast of Sharon Muhlenburg County, Lovell Cave Grove, Todd County. One entrance, represents what is probably the a resurgence, is at the head of a largest cave in the county. It is branch of Antioch Creek, 1.6 km as rich in history and local southwest of Antioch Church. The legend as it is in passage. During other entrance is a sink (karst the 1920s, a man named Clark ran window) 1.1 km a1most due west of , Love1 1 Cave as a local tourist the spring mouth. A stream emerges attraction. Tours through the from the western side of this sink walking passages of the cave and runs into the cave via the averaged 50 cents. On warm spring back entrance. The cave has 1.49 Sundays when the cave was open for km of passage and runs almost business, hundreds of people straight west to east (Fig. 5), camped on the ridge above the para1lel to prominent east-west entrance. By 1928, the normal faults in the area. commercialization of Lovell Cave The resurgence entrance opens had ceased. in a headwall at the western end of a ravine and is 12 m wide by 3 fictually, the cave's history m high. A walking passage with dates from much earlier. Indian ponded water trends initially artifacts have been found both southwest for about 30 m. Here it around the entrance and within the doubles back to the north as a cave. The walls of the cave carry narrow walkway, with waist-deep several smoked dates in the 1800s pools, for 45 m, emerging over including "J. K. East 1833" and a breaddown into the ma.in part o+ dubious "Jessie James 1868." Local the cave. The entrance streamway legend also fills the cave with may be partially bypassed through runaway slaves and an improbable a dry.upper level crawlway. At the number of bankrobbers. The most room where these levels rejoin, unlikely story told about the cave the stream strunk assumes a nearly is the story of the discovery of due-west trend, which it maintains several "Egyptian " that +or approximately 1 km, nearly to were supposedly viewed by Floyd the back entrance. Collins himseif. The interior of GLOVER'S CAVE the cave has suffered from tourism right angles. With the exception to the extent that even 10-m-high of White Bat Passage on the ceilings bear only the stumps of eastern side of the cave, a1 1 broken stalactites. In many places these passages are tied together. smoked or painted names obscure A moderately large chamber on the the walls. western side of the passage is Lovell Cave (Fig. 12) opens tied to the other passages only by through a 5 m wide, 2 m high the extremely tight Hard Way Tube. entrance in the southern side of a As Tourist Ovenue continues north, ridge. 6 passage with a typical the side passages intermesh in a arched shape leads down a fairly more and more complicated system. steep slope into the entrance Several passgeways are yet to be chamber of the cave. The entrance pushed to their fullest extent. chamber is in the sandstone Over 1 km of passage has been caprock of the cave, with a floor surveyed. composed mostly of large breakdown blocks. Two passages lead from the Glover's Cave left (western) side of the The cave entrance is located on entrance room and continue west the eastern bank of the western until they meet Tourist Canyon, a fork of the Red River in long canyon passage going southeastern Christian County north-south. The southern end of (Fig. 8). From the Primary (or Tourist Canyon terminates in Main) Entrance, the cave proceeds breakdown, but a narrow stoopway north and east. The northern branches west and enters a wide branch begins as a wide tube 12 m crawlway that contains an by 3 m, and 45 m from the entrance intermittent stream, flowing runs into a cross passage trending southeast to northwest. As are all east-west. To the west, the cross water features of Love11 Cave, passage, an oval tube 8 m wide and this stream is extreme1y variable. 1 to 2 m high, ends in sediment A series of small interconnecting and collapse. To the east, the passages exists in this area. Back tube runs 110 m, with much mud, to at Tourist Canyon, moving north. a terminus. At the junction with passages frequently meet at nearly the cross passage, the northern branch changes to dimensions of 3 to 4 m wide and 4 to 8 m high. This branch runs to the north for LOVELLCAVE 75 m to a chamber with a covered MalLENBURG COWTV, KENTUCKV mwmnn sl'm1cAL suRnr 10-m pit in the ceiling (with water pipes coming down) and a short walkway to the surface through the door at the Secondary Entrance. North beyond this chamber is a complex area of WHITE BAT PASSAGE entrances, pools, and siphons. This area is used as a water N TOURlST CANYON 5uppl y. Trending east from the Primary Entrance is the main passage of ENTRANCE 0 feet 100 CHAMBER the cave, called the Lincoln . This passage runs for 600 m with remarkably uniform ceilings. At many localities in the passage are rimstone dams. After 600 m, the passage forks.

LNlRANCL The left fork trends northeast for 90 m as a sediment and breakdown Figure 12. Lovell Cave, floored passage 3 to 6 rn wide and WESTERN KENTUCKY REGION 135

1 to 3 m high. This Lake Passage ends in a deep lake in a room 18 m long, 10 m wide, and '3 m high. The Lake has a definite water flow from the southeast to the northeast and may be part of the water seen at the resurgence. The pool has abundant cave life plus many leaves, sticks, etc., embedded in its shore. From the incoming passage to the lake is a downward 6 m slide. The right fork is a continuation of the main passage. Thirty meters from the fork, a stream is met that cuts to the northeast under the passage wall into a short side pa5sage leading to a domepit in a decorated chamber called the Grotto (Fig. 13). The stream sinks here, and is believed to go to the Lake. only 60 m to the north. Just beyond the passage leading to the Grotto is another side passage that leads east 210 m as a walk. crawl, stoop-walk, and terminates just beyond a small dome in the ceiling. This passage was recently ex tended. The main passage, now carrying a stream, meanders south 200 m as a large tube up to 13-m-wide and 8-m-high. ht the Dining Room. collapse blocks the passage. A Figure 13. Formations in the side lead at this point, the Grotto, Glover's Cave, Christrian Winder, trends 200 m west. The County, Kentucky. main passage collapse can be forced to regain the major trend and a short passage, the Ballroom. Christian County (Fig. 8). Over 5 The next 750 m alternates between km of passage has been mapped, and solution tube, breakdown passage, survey work continues. The Main and breakdown crawls, averaging 3 Entrance to the cave is a to 6 m high and 8 to 15 m wide, spectacular double level entrance. passing the Eagle, the Guillotine, The upper part is a large shelter and the Cliff Hanger. The main that becomes a small passage passage ends in a formation area. ending after 60 m. Two passages the Beauty Parlor. f4 further low, lead off from each side of the wet, and muddy passage can be entrance. The northern passage is followed to an apparent end at a an overflow tube of most1y walking siphon, domes, and sediment size, with some ponded water. It blockage. Over 3 km has been lowers to a crawl then reaches the surveyed. surface on the banks of the river 160 m upstream of the Main Twin Level Cave Entrance. Twin Level Cave is located on The major portion of the cave the eastern bank of the West Fork leads south from the right side of of the Red River in southeastern the Main Entrance. An easterly 136 BIG SULPHUR CAVE crawl of 100 m leads to a junction trickle tributary emitting the with the main stream. Downstream sulphur smell. A more important to the left, the main stream junction is 30 m past the "sulphur passage sumps immediately, but crawl, " also on the right. This upstream to the right it continues junction is a climb-up to an upper 2 to 3 m high and 2 to 5 m wide level section, the Bat Passage, for 350 m to a collapse sinkhole which paral lels the main stream entrance, Turner's Window. Pushing passage for over 200 m. It upstream the passage continues as averages 5 m wide by 2 m high. before, with some breakdown and At a point bB0 m from the low crawls, for a further 1,500 m entrance, the stream forks into to another entrance at Carneal 's segments of about equal volume. (Watt's) Cave. The cave continues The right branch continues to most upstream another 1,000 m to an of the known cave as the Middle additional entrance, Ham1 et 's Cave section. At the fork straight Well, a drilled well 1.2 m in ahead (southeast) is a narrow diameter. Total survey to date is walkway with ponded water. This 5 kin, with leads still remaining. walkway appears to aburptly end after 30 m, but beyond a low Big Sulphur Cave near-sump, this passage, the D Big Sulphur Cave, 4.8 km long, Survey, continues 150 m southeast is situated in southeastern Trigg to where a large block halts County on the bank of Little River progress despite airflow. Right at the apex of a sharp meander. A from the major fork is the way to distinct, though not overpowerinq, the Middle Cave and most of Big sulphurous smell is noticed at the Sulphur. The initial trend remains entrance (Fig. 14). southward but the main passage The first few meters of passage soon assumes a sharply different is a stoopway over small direction than the entrance trunk: breakdown. Thereafter, one follows west or northwest, roughly a south-southeast-trending stream paral 1el ing the surf ace course of channel of mostly rectangular Little River, but with an opposing cross section, averaging 4 to 7 m stream-flow direction. This section width and 2 m height (a few short averages 5 to 10 m in width and stretches require easy crawling). 1.2 to 4.5 m in height, with two The stream generally occupies the or three slightly offset levels in whole floor but is usual 1y no more most places. than ankle deep. The first lead on On1 y 25 m past the mainstream the right is 75 m from the fork to D Survey is another entrance, a 20 m crawl with a significant intersection; on the right, several meters above stream 1eve1 and not obvious at first glance, is a junction of an upper level. This is a truncated segment of the Bat Passage described ENTRANCE above. It begins as a wide but low crawlway, changes to a sizable V breakdown route similar to the Bat Passage, trends northwest for 185 m, and ends in collapse very close to the southern end of Bat Passage. After another 35 rn south, the F Survey trends southeast for 225 m, averaging 1.5 m wide and 0.5 to 1 m high. It "ends" by breaking up into several small components that Figure 14. Big Sulphur Cave. are not negotiable. The F Survey WESTERN KENTUCKY REGION passage close1y para1lels the D Cool Spring Cave Survey at a slightly higher level; Cool Spring Cave is located on it probably carried the same water the Caledonia 7.5+ minute at one time. Both these passages topographic map, on the northern are close to a nearby short cave, bank of Sinking Fork in central Little Big Sulphur Cave. Trigg County. The cave is complex The cave swings west, and just and the second largest known in over 100 m past the F Survey western Kentucky. Unusual for the intersection is a wide but area, Cool Spring Cave has somewhat low junction of another extensive upper l eve1 development , left branch, the H Survey. This with three distinct horizons. Some semi-maze, broken-dawn area trends good leads remain and it is hoped west-southwest, varying from the cave will yield more virgin walking to easy passage in the future. Over 5 km crawling size and totals about 280 of passage has been surveyed m. The N Survey, is also on the (Figs. 4 and 15). left, 120 m after the H Survey. Cool Spring Cave has three This section, a 270 m loop, is known entrances, a1 1 grouped surmised to be a portion of the around the resurgence. The Main cave's principal but intermittent Entrance, carrying the cave upper level. It begins as a stream, leads north and then west south-trending walkway for 50 m, 145 m as a broad oval tube 8 to 10 then a tight crawlway continues, m wide and 1.5 to 3 m high. The doubling around to the northwest sinkhole entrance enters through and rejoining the main trunk 80 m the ceiling of an alcove of this beyond the initial N Survey passage just in and to the west of junction. the entrance. At 145 m, the stream There are no more significant leads off the main passage for the succeeding.320 m after the second N Survey junction, until a breakdown zone is encountered. One can stay low and right, with the stream, skirting the breakdown into the mainstream continuation to the west. A slightly more obvious choice is to crawl straight ahead into the breakdown. Ey doing so one pops up after 20 m COOL SPRING CAVE into the Big .Rooms area. The first room in the Big Rooms area is TRIGG COUNTY, KY roughly rectangular in cross SOUTHWEST KENTUCKY sections, approximately 20 m on STUDENT GROTTO each side and 5 m high. Thre are WESTERN KENTUCKY three principal directions from SPELEOLOGICAL SURVEY this room. Near where the L Survey departs the Big Rooms section are two CRYSTAL FLOOR PASSAGE interconnecting and inconspicuous \ L- but significant side passages on two levels, the lower with a small stream. 6 meandering crawl only 1 m wide and 0'.5 m high trends west for 325 m, ending in several mud slumps. This is the fartherest

point from the entrance in Big DRY ENTRANCE Sulphur Cave, a 1i ttl e over 1,500 m. Figure 15. Cool Spring Cave. DECIBEL CAVE passage opens into a large Junction. From this point, the chamber, Pathos Parlor, 90 m long, middle level extends north 60 m as 15 m wide, formed by the a tube 3 to 6 m wide and 1 to 2 m intersection of the three levels high to Junction, where the of the cave. To the left (south) , lower level stream passage is it proceeds to a passage that ends rejoined. The middle level in a mud. flowstone, and breakdown continues north across the stream choke in the upper level. Just as a walking passage, gradually before the choke, a passage leads diminishing to a crawl and ending down and east as a low, wide stoop after 120 m. The stream passage and crawl to the Dry Entrance, as siphons downstream to the a segment of the middle level of southwest, but continues north as the cave. Crawdad Crawl for over 600 m, From the northern end of the gradual 1y lowering from a walkway Pathos Parlor, passages leave at to a stoop to a crawl. It a1 1 three levels. The upper level continues, upstream from Guano leads west as a tube 4 to 8 m wide Junction, and the middle level is that starts out 2 m high but intersected again. It proceeds dimini shes, final 1y ending in east 180 m as a wide, low tube, sediment fill after 210 m. Leaving Crystal Floor Passage, and north the northern end of Pathos Parlor, over 1,100 m a1so as a wide, 1ow the stream or lower level trends tube, Helectite Pass. Its end has northwest for 200 m to a sump as a not been reached, as it tube 1 to 2 m high and 6 to 9 m degenerates into a very low crawl, wide. Sixty meters from Pathos Wide Way. A sharp right from the Parlor, a large breakdown occurs. base of the Guano Junction leads Up and to the south through the as a diminishing tributary breakdown is a connection to the passage, Mud Cave, for 500 m to upper level. Just before the sump the east. the lower level crosses under the Cool Spring Cave is a very middle level to form a large room. complex cave, and good potential To the east this middle level exists for significant extension 1eads past an important side of the cave. The source of the passage at Breathing Junction and water in the cave is not known, continues over 300 m, connecting and the major tributary passages into Pathos Parlor from the east have not been reached within the and into the Dry Entrance area. cave. The plan of the cave To the west from the siphon suggests that it developed area of the main stream, the initially by gathering water from middle level runs 240 m to the the Sti1 lhouse Branch tributary of southwest, as an oval tube 3 to Sinking fork. Much of the current 4.5 m wide and 1 to 2.5 m high water flow in the cave is believed ending in the Bat Room, a large to originate from further upstream chamber complex formed by in Stillhouse Branch rather than intersection with the upper level. in past stages of the cave's For the next 300 m the middle development (Fig. 4). level is criss-crossed by the upper level many times, resulting Decibel Cave in considerable passage complexity Decibel Cave, containing almost and over 900 m of passage. Ample 2 km of passage, is located on the wind and many small domes indicate northern bank of Sinking Fork west a possible entrance nearby. These of Mill Stream Spring (Fig. 4). passages end in sediment fill, The entrance is at river level in flowstone, or collapse. No active a 1ow li mestone bluff , and i5 1 m streams are seen in this section high and 2 m wide, with a stream of the cave. flowing out. This passage leads The major portion of the cave north into a junction room. To the lies north of the Breathing east, a dry crawl at roof level or WESTERN KENTUCKY REGION a wet crawl at floor level unite The most easterly of the passages after 15 m in a low, wide stream has a short tube 20 cm in diameter passage, continuing west 20 m that leads to the surface, but it before opening out into a large is not negotiable. breakdown chamber over 30 m across The cave is isolated in a low and up to 6 m high. The stream hill that extends to the southwest comes out of impenetrable between Sinking Fork and Steele breakdown at the extreme northern Branch (Fig. 4). The cave point of the room. apparently conducted water from From the junction room near Steele Branch through the hill to the entrance, a low, muddy crawl Sinking Fork, as evidenced by can be followed north into a large scalloping in the passages in the corridor. The main corridor swings western and northern portions of southwest from these side passages the cave. The water seen in the and the cave suddenly becomes very eastern portion of the cave is complex. From the broad, low room believed to be derived from Steele with several bedrock pillars, a Branch, but this has not been short crawlway leads north to a proven. dome. To the west a series of tubes lead, after a short jog Twin Tunnels Cave south, to a large chamber with Twin Tunnels Cave is located on many -bedrock partitions. Short the southern bank of Sinking fork side passages 1ead east-southeast in Trigg County (Fig. 4). Steep from this room, but to the bluffs line the southern bank of west-northwest a group of para1lel the stream at this point, and in passages open into a very large this bluff, 3 m above stream chamber floored with breakdown. level, are the two entrances to Fassages radiate in all Twin Tunnels Cave and some other directions, and the room seems to minor associated cvaes. have gained its breakdown and The two entrances are 30 rn broad expanse due to the collapse apart, and lead to passages 1 to 2 of many bedrock pillars. Leading m high and 4 to 8 rn wide. The north out of this room is a low (1 passages join 80 m into the hill. m), wide (3 to 6 m) tube that From the junction, the passage winds for 150 m north, west, continues west as an oval tube 4 north, east, and then north again to 8 m wide and 1 m high, with a ending in a breakdown choke. Some mud floor continuing a meandering high rooms are passed along the trench 1 to 2 m deep for 300 m, way, and numerous vadose dome pits where a very low and wide side esist. The wal Is of this tube show passage winds northeast for 30 m. we1 1 developed scal 1oping , This passage is believed to be a demonstrating past flow to the tributary passage associated with south and southwest. some large, blocked entrances on the bank of the Sinking Fork, 60 m From the large central chamber. downstream of the main entrances passages lead both west and south. to the cave. To the west, a series of parallel tubes run for 80 m before ending The cave jogs briefly south in dome pits, mud plugs, and be+ore continui ng west, breakdown. These passages have encountering a complex junction scalloping that shows past flow to area 450 m from the entrance. Very the east. To the south, a series wide, but very low passages lead of low arches trending north and south out of the room. south-southeast intersect larqe Like the previous side passage, a tubes oriented along joints and thin person could perhaps gain trending west-northwest and more cave here. The meandering east-southeast. These,cross trench cut in th floor continues passages end in mud, dirt, west 40 m, emerging after a short breakdown, and flowstone plugs. climb, into a large room 50 m HUSK (LAWRENCE) CAVE long, oriented north-south, with their in southeastern its western wall composed of Caldwell County. The water is massive breakdown. This room plots seen in Harmony Spring Cave, a out as being under a large, complex 1,404 m system that has closed-contour depression, the four entrances to a major river cave having penetrated a ridge passage with overflow routes and between the active Sinking Fork tributaries. It sumps upstream and and an abandoned meander. downstream at major karst window Ancient and present day flow entrances. Harmony 'Church Cave can markings show waterflow into the be entered 200 m south of the 580-m-long cave. The cave Harmony Springs downstream apparently functioned as a meander entrance. Upstream it sumps very cutoff cave, and was abandoned close to the downstream sump in when its point of discharge became Harmony Spring Cave. It continues a sediment-filled oxbow of Sinking downstream past a collapse fork (Moore and Mylroie, 1979). sinkhole entrance to 2.5 km of passage, mostly very wet stream Husk (Lawrence) Cave passage or oval flood-overflow routes. The water is next seen Husk (Lawrence) Cave is located 2 km to the southeast at the in northern Trigg County. The main upstream sump of Perkins Spring entrance is at the eastern end of the prominent 300 m long karst Cave; then it flows through a very window in a broad sinkhole plain. deep river passage for 300 m to The mouth, 18 m wide and 8 m high, the first Perkins Spring Cave drops down a mud slope to a trunk entrance at a karst window. carrying a stream. Downstream, the Continuing downstream, nine more water immedi ate1y vani shes into an entrances are located before the apparently impenetrable breakdown water resurges 150 m choke, emerging from a secondary south-southwest in a karst window. entrance in the middle of the The water crosses the window and sink, and flowing underground sinks again, reappearing at again at the western end of the Martin's Spring on the eastern sink. bank of Kenady Creek less than I km south--a classic sinkhole plain Upstream from the main entrance cave with almost no abandoned is a passage, 6 to 12 m upper 1eve1 s. wide and 3 to 10 m high. Large silt banks continually line the Skinframe Sinks (Rice) Cave stream on one or both sides. The trunk trends predominantly The main entrances are located east-southeast, with a couple o+ in west central Caldwell County in major meanders, for an estimated an almost vertical-sided, 1,200 m to a deep pool. Beyond, el1 iptical collapse sink, the character of the cave changes, approximate1y 30 m lang by 15 m becoming much lower, with the wide and 4 m deep at the terminus stream occupying the entire floor. of Ski nf rame Creek's normal 1y dry Seventy-five meters beyond the bed (Fig. 16). pool, it becomes necessary to From its entrance, the main stoop, with only a few centimeters upstream passage has been mapped of air space. The cave may well for 1.6 km south and east, with a continue for a substantial major branch to the southwest. At distance, however. 225 m upstream from the entrance, the stream forks, the lesser flow Harmony Church Cave System coming from a major right branch, The Harmony Church Cave System the Brewster Creek section. consists of three major segments Beyond, the main upstream passage totalling over 4 km of passage. continues, now predominantly Two major streams, Millwood Creek east-southeast. fit 590 m from the and Battle Creek. sink close to entrance, the stream is lost at a WESTERN KENTUCKY REGION

Brewster Creek sinks. The downstream main entrance passage opens to a 8 m wide. 1 m hiqh overf 1ow trending north-northeast. At 80 m, the main stream joins from the upstream portion of the cave. The passage beyond was explored 250 m, trending north and averaging 6 to " 10 m wide and 1 to 2 m high to an apparent sump. The stream throughout this branch is ponded into one large "lake," averaging 1 Figure 16. Skinframe Sinks Cave. m deep, and usually occupying the entire f loor. sump, and further progress is by Cross sections of Skinframe way of an overflow route. At 650 m passages are typical of western the only relatively large room is Kentucky sinkhole plain systems: seen in this branch (about 15 m uniformly arched, 3 to 5 times long, 8 m wide, and 2.5 m high): a wider than high. Survey data small stream is seen here but can indicates portions of the cave not be f 01 lowed in either near the entrances to be generally direction. The continuing passage, 10 m or less beneath the surface, an overflow route the entire although the stream extension may distance, is chracteristic of the have 15 m or so of overburden. A previous1y explored cave: a few unremarkable spel eothems in generally damp-dry the Brewster Creek branch were the gravel/cobble-floored elliptical on1y f ormati ons observed. Large tube subject to complete flooding, , conspicuous in over-Flow averaging 3 to 5 m wide and 0.5 to pools, are common. The total 1 m high, with one brief stretch surveyed extent is 3 km. of walking passage and a few other stand-up spots. Mill Bluff Cave A1 thobgh a small tributary The 1ocall y we1 1-known opening stream occupies the extension to Mill Bluff Cave is 2.6 km passage for 100 m, the main south-southwest of the town of Skinframe Creek is not seen. At Fredonia. The entrance, in an Procrastination Room, 1,500 m from overhanging cliff of Ste. the entrance, is a fork; the right Genevieve 1imestone, is 1i kel y the branch, of standing height and largest and most striking of any occupied by knee-deep, stagnant, cave in western Kentucky, being 50 debris-laden water, goes 60 m to m wide and 18 m high. an apparent sump. The other way, a Three spacious passages lead rather low crawl over silt beds, back southeast from the entrance, goes 140 m to a low room and two of them are f 1ood-overf lows, branches, both ending within 50 m. and all unite within 200 m of the The south-southwest-trending entrance. A major stream passage, Brewster Creek branch of the B to 14 m wide and 3 m high (1 m upstream cave has been surveyed of water) continues a further 200 over 700 m from its intersection m to a complex junction area. with the main stream. Passage North is a large passage with deep dimensions average 3 to 4 m wide water sumping a+ter trending and 1 m high. At 425 m from the north-northwest for 160 m. junction is a room 15 m long, 11 m Continuing east and then northeast wide, and 3.5 m high, with three is a major stream passage, sumping openings to the surface. This after more than 550 m of entrance complex is in a small additional passage. A series of sink just northwest of the main overflow routes parallel the 142 LISANBY CAVE stream passage to the south, by Mrs. Lisanby's father, James W. joining both at the junction Hollingsworth, who died in 1915. complex and near the terminal The owner married the late Alvin sump. These passages alternate Lisanby, an attorney, in 1921, and between wide, roomy tubes and they remained on the Hollingsworth muddl y crawls. The cave, believed farm. There may have been no to drain a large area of Caldwell reference to this cave in County to the east, penetrates speleological literature prior to over 1 km eastward and contains 1962-63. 2.4 km of passage. The main entrance (6.5 m wide and 1.0 m high) to Lisanby Cave is Lisanby Cave in a collapse sink that intersects a fossil trunk passage. From the The main entrance to Lisanby entrance, one can go either left Cave (the only one verified prior (upstream) or right (downstream) to 1983) is located on the (Fig. 7). The main Left Branch southwestern outskirts of the town passage is a paleo-condui t, now of Princeton, county seat of dry, with thick clay fills. It Caldwell County. The cave is meanders upstream in a general 1y locally famous and no doubt has northwest to north direction and been famous for a century and a is characteristically 4.5 to 12 m half. However, to date only wide and 0.6 to 3 m high. In sketchy information on the cave's places, a small stream is history has been gathered. The accessible in a 1ower-1 eve1 current owner, Mrs. Alvin Lisanby , crevice passage, sometimes indicates that it was previously unroofed by an overlying passage. referred to as the "Saltpeter Cave"; it may also have been known On the left, 200 m from the entrance, is a rather obscure in the early 20th Century as junction of a significant side "Hollingsworth Cave" after Mrs. passage, Thunder Road, which Lisanby's father, who was a continues west for 180 m, ending Confederate soldier. Although in breakdown. Its stream may come heavy contemporary traffic through from Watson Cave, 1.5 km to the the main part of the cave seems to west. The main passage continues have obliterated most if not all north, with a few minor side evidence of saltpeter mining passages, for 550 m as a gradually activities, there is a good chance lowering trunk passage, ending in that the cave was mined during the col 1apse. Civil War period. The relatively dry passages near the entrance The Right Branch route is (especially the right branch) similar to the Left Branch superficially appear conducive to described above: it is a generally nitre production. Entrenchment in dry, clay-f i1 led paleo-trunk. It a passage between the entrance and estends southeast for abut 435 m first room in the right branch may to apparent1y terminal breakdown. have been the work of miners. It consists of a chain of walking According to Mrs. Lisanby, height rooms connected by easy deserters (from which army?) hid crawlways. Among the often in the cave during the War Between multiple options, the small stream the States. Signatures dating to from Thunder Road in the Left the 1870s may be seen in Lisanby Branch can be reached in places. Cave but most historic Besides being the most probable inscriptions have probably been site of saltpeter mining in obscured by the great volume of Lisanby Cave, the Right Branch is modern defacement in the parts O+ also the most heavi 1y visited the cave near the entrance. The portion of the cave, with many owner also recollects dances held hundreds of inscriptions and in the cave when she was young. usually a good deal of litter in The property was acquired in 1900 evidence. WESTERN KENTUCKY REGION

Just before the dry trunk water more than ankle deep is seemingly ends in breakdown, the encountered on1 y in a few pools. way onward is a 4.5 m climbdown to The passage meanders considerably a stream passage, i. e. , the water but the basic trend is first seen in Thunder Road. This north-northwest for over 1.5 km. route trends southeast as a walk Until almost the very upstream end and scramble, of ten on two levels. of this segment, there is only a Only 30 m from the climb-down is a single minor side passage, a keyhole or which discouraqes tight, 100-m-long tributary from nearly all "flashlight cavers" the left at a point 1.1 km (judging by the lack of vandalism upstream from the sump. beyond). Just over 200 m from the An abrupt change takes place climb-down is an easily unnoticed 1.48 km upstream from the sump as lead at a slightly higher level. several connections occur with a This is the beginning of the rather extensive complex of dry Farless Extension described in the upper levels. Just beyond is the following section. In the next 100 Poison Gas Room, a breakdown m is a side passage containing a chamber 6 m wide and high with segment of the Farless Cave three ways on: a connection to the stream, and a second keyhole or is dry upper levels; the low encountered, on the other side of continuation of the "river", and a which is the rather surprising 117 m long tributary entering over intersection with the cave's main a small rimstone cascade on the stream, Lisanby River, at a point right (north) , cal led Waterf a1 1 785 m from the entrance. Trai 1. From the obscure connection to Something oi a maze of dry main Lisanby noted in the passages is developed above preceding section, the Farless Lisanby River around where it Estension begins as a tight, dry, reaches the Poison Gas Room. These twisting crawl trending south then passages are on two sligtly offset west: at 50 m, a paleo trunk levels, 3 to 6 m above the stream passage is intersected. This and vary from easy crawlways to passage averages 4 or 5 m wide and barely wal kable dimensions. In a1 1 1 m high. One direction is likelihood, the upper levels above west-northwest for about 100 m the Poison Gas Room--as well as before ending. The other way is those farther upcave--were southeast. rather wide, low, and original1y integrated with the dry; it leads to a series of upper Left Branch section complex but level fragments and lower level now are interrupted by collapses. stream crawls. Over 700 m of The various upper level segments variable passage leads to the in the Poison Gas Room area climbable entrance drop of Farless comprise some 838 m of passage. Cave. Upstream (west-northwest) of Exact1y who discovered the the Poison Gas Room, the main Lisanby River, a very significant Lisanby River conduit is reduced part of Lisanby Cave--and when--is to a rather unappealing crawlway, uncertain. The Lisanby River trunk referred to as the Canal or Elbow is remarkably uniform in cross Runway (the 1atter name referring section and nature, varying from to grooves in the little m~rdbanks 3.5 to 6 m in width and 2 to 3.5 m on one or the other side that in height. This conduit in massive cavers wear while traversing the Ste. Genevieve limestone has few crawl ) . This crawlway is 3 to 4 m features to speak of, besides wide and somewhat under 1 m high, modest-size silt banks and with ponded water, for a distance scattered solutional oddities. The of some 125 m. Midway is a stream carries about 0.03 m3 distinct1y lower cei 1ing point, per second of water in normal +low normally no problem but which and has just enough gradient that could be closed after high runoff. BAKER CAVE

On the upstream side of the the banks of Eddy Creek. Along the Canal, the streamway opens back up way it picked up tributaries from to barely walking or stooping the west and southwest at Trilevel size, not quite as spacious as Crawl , Thunder Road, and Far less before the Poison Gas Room. One Cave. The lower levels of these hundred forty meters farther, a three tributaries provide water to significant fork of the stream is the Lisanby River today. reached, The Y. About two-thirds Total passage length is 11.3 of the flow is from the left km, with some possibilities for (western) branch and one-third extension. This makes it more than from the right (northern) branch. twice as long as any other known The northern branch leads to.a Western Kentucky cave. series of stream passages, first Dungball Crawl and then Waterfall Baker Cave Canyon, trending to the northwest Baker Cave is located 4 km over 500 m before becoming north-northeast of the village of extreme1y low and wet. The stream Salem in Crittenden County. The passage intersects at two widely cave is developed in the Renault separated points along its Limestone, f 01 1 owing joints, southern wall, a dry upper level faults, and veins striking complex, Lost Watch Passage. This north-northeast that are related passage trends to the nearby Babb and Levias northwest-southeast, becoming fault systems.. The cave has been extremely low at its northwestern mined for fluorite, and shot end, but in the southeast holes, a collapsed shaft to the direction it swings south over the surf ace, and various mining left branch of The Y, connecting paraphernalia can be found in the with further upper levels that cave. continue both south and southeast. The cave has two entrances, Interconnected upper levels, the situated close to each other. One Tri1 evel Crawls, join the Lost is a crawlway off a small sink on Watch Passage and wander over the a hillside that goes west about 7 left branch stream to where it in m then intersects a shaft to the turn branches. The southwestern surf ace (the second entrance) branch contains a small stream in about halfway down its 8 m depth. a multilevel canyon complex called A climbdown reaches the base of Sandpaper Canyon that becomes too the shaft which slopes immedi ate1y small in less than 200 m. North into the main stream passage. from the Sandpaper Canyon junction Upstream, to the north, the overflow passages reach Zig Zag stream passage starts out 3 to 5 m Canyon, a stream canyon that leads wide and 1.5 to 3 m high. A fault west with over 1 km of passage or joint is prominent along the before ending in mud plugs and eastern wall. After about 50 m, sumps. The upper 1evel passages the passage becomes less than 1 m associ ated with the Lost Watch high, half full of water, but Passage and Tri1 evel Crawl total still 3 m wide. The passage has over 2.5 km of passage. not been pushed beyond this point, The gross morphology of the although it appears to continue. cave w~uldsuggest that it began Downstream, to the south, the as a major stream passage carrying passage is initially easy walking water from the upper and scrambling for about 40 m, 2 (northwestern) end of the Lost or 3 m high, 3 to 5 m wide. The Watch Passage southeast past the passage then degenerates. with the current position of The Y to the major upper level being blocked by upper levels off the Poison Gas sediment fill or collapse, and Room and on through the Left and progress is restricted to a 0.5 m Right Branches of the main high crawl in the stream. This entrance series to a resurgence on crawl opens up after 25 m into a WESTERN KENTUCKY REGION 145 large chamber f i1 led with mining WKSS, welcome interested, debris, and containing a collapsed conservation-minded cavers to mine shaft on the western side. come, locate, explore, and map From this point downstream the caves with the WKSS. passage becomes extremely pleasant, being mostly walking height in a fine tube up to 6 m REFERENCES 1 wide 'and 4 m high. Flfter crawling over and around an occasional Carstens, Kenneth, 1980, Savage breakdown pile, a massive terminal Cave--The future of its prehis- collapse is met an estimated 500 m tory, In Mylroie, J. E., ed., from the entrance. One short 20 m Western Kentucky Spel eological side passage was noted. Survey Annual Report for 1980, Murray, Kentucky, p. 17-28. CONCLUSIONS Dever, G. R. and McGrain, Preston, The Western Kentucky Coal Field 1969, High-cal cium and 1ow- Region is a poorly explored but magnesi um li mestone resources in extensively cavernous, low re1ief the region of the Lower Cumber- sinkhole plain. With almost 100 km land, Tennessee and Ohio of passage surveyed in less than Val 1eys, Western Kentucky: Ken- 10 years, the continuing potential tucky Geological Survey, ser. of the area is exciting. Because 10, Bulletin 5. 192 D. of the region's relative obscurity in speleology, it has provided an Mason, D. , MrDowel 1 , D. , and opportunity for a Mylroie, J. E., 1984, Meander conservation-oriented, scientific, cutoffs in the Glover's Cave and comprehensive survey area, Christian County, Ken- organization, the Western Kentucky tucky, Mylrqie, J. E., ed., Speleological Survey, to explore, Western Kentucky Speleol ogical study, map, and compile cave data Survey Annual Report for 1982 free from complication and and 1983: Murray, Kentucky, p. poli tics. A number of diverse 7-10. groups and indi vi dual s work together in cooperation with the McFarlan, A. C., and Jones, D. J., WKSS to maximize both the study 1954, Geologic map of Kentucky: and use of the region's caves. Kentucky Geological Survey, ser. Extreme efforts have been made to 9, Scale 1 in. = 16 mi. maintain excel lent 1andowner relations and to preserve the Moore, F. M. and Mylroie, J. E., caves (Fig. 13). 1979, Influence of master stream A1 1 caves are valuable and incision on cave development, priceless. The value of the Trigg County, Kentucky, in, western Kentucky caves depends on Mylroie, J. E., ed., Western the observer 's point of view. Kentucky Speleological Survey Cavers interested in deep pits, Annual Report for 1979: Murray, endless spacious galleries, and Kentucky, p. 47-68. fantastic mineral displays will be disappointed. Cavers from areas Whaley, P. W., and black, P. C., with small caves, from cold 1978, A brief geological de- climates, or areas with vandalized scription of the Western Ken- and heavily traveled caves will be tucky survey area, in, Mylroie, pleased with most western Kentucky J. E., ed., Western Kentucky caves. The authors of this Spel eol ogi cal Survey Report for chapter, as representatives of the 1978, Murray, Kentucky, p. 5-10. Chapter 9 CAVE LIFE OF KENTUCKY Thomas C. Barr, Jr. Department of Biology University of Kentucky Lexington, Kentucky 40506

The unusually rich life of Ken- nating bat colonies, but describ- tucky caves reflects the fact that ed nothing else. The eminent large segments of the two major Russian coleopteri st, Baron T. Missi ssi ppi an 1i mestone plateaus Victor von Motschulsky, attracted of the eastern United States are by Tellkampf's descriptons of found within the Commonwealth. blind , visited Mammoth A1 though siqni f icant but 1i mi ted Cave in 1854--no small achieve- cave faunas occur in the Flue ment in those days--and described Grass and in isolated karst additional species he found there. islands along the Rough River A few years later Alphaeus Spring Fault Zone and the northwestern Packard, Jr., taking advantage of face of Pine Mountain, the Missis- a free trip to Mammoth Cave offer- sippian Plateau caves harbor the ed to participants at an diverse troglobi te faunas for Indianapolis meeting of the Amer- which Kentucky is best known. ican Associaion for the Advance- These faunas are as rich as those ment of Science by the L & N Rail- of any other karst region in the road, collected still further cave world. inhabitants. Dissatisfied with Historically, Kentucky bio- Darwin's theory of natural selec- speleol ogy began when Stephen tion because he believed it did Bishop crossed Mammoth Cave's not entirely explain evolutionary Pottomless Pit on a crude cedar phenomena, Packard subsequently ladder and discovered Echo River visited Mammoth and other caves, and its blind . When including the Carter Caves in James DeKay (1842) described northeastern Kentucky, as well as Ambl~oeq&~spelaea in a footnote Wyandotte Cave, Indiana, and in his "Zoology of New York", a Nickajack Cave, Tennessee. He german physician, Theodor mistakenly believed that life in Tellkampf, visited Mammoth Cave in caves exerted a direct effect on search of specimens of this, the the evolutionary loss of eyes and world's first known blind cave- pigment and was for a time a cham- fish. He a1so collected and de- pion of the Neolamarckian school^ scribed blind beetles, a blind of thought ( see Packard, 1890: harvestman, a blind , and a Barr, 1968a). Packard's contribu- blind crayfish (Tellkampf , 1844a, tions, though leaving much to be 1844b, 1845). Constantine desired in the way of accuracy and Rafinesque (1832) had visited heavi 1y 1aden with hi5 Neol amar- Mammoth Cave almost two decades ckian philosophy, at least pub- earlier and observed large hiber- licized the rather rich fauna of CAVE LIFE OF KENTUCKY

Kentucky caves ( see Packard, (1966). The inventory stage is 1888). still going on; additional trog- When Carter Hovey ex- lobite species are being dis- plored Mammoth Cave he was at covered and described, but at a times accompanied by R. E. Call, 51ower rate than previously. who wrote a chapter on cave life Studies of bats in Kentucky in Hovey's famous guidebooks to caves reached a peak in the 1960's Mammoth Cave, published in sever- and early 1970's with the work of al editions around the turn of the R. W. Barbour, W. H. Davis, and century. Northwestern University's their students at the University Professor Or1 ando Park (a native of Kentucky. This research cen- of Eliiabethtown, Kentucky) took tered primarily on hibernating his classes to Mammoth colonies in Mammoth Cave National Cave in the 1930's and actually Park, Carter Caves State Resort published a key to Mammoth Cave Park, and a small number of caves animal species as an exercise in in Lee County (Davis and Barbour, his laboratory manual on animal 1965; Hassell and Harvey, 1965). ecology and (Park, 1939). For consideration of cave The key is now considerably out- faunas, a regional classification dated, but it includes the species slightly different from that used described by Call, ReKay, elsewhere in this book has been Motschulsky, Packard, Tellkampf, employed , and reasons for organi z - and others, as well as observa- ing discussion around this scheme tions and redescriptions made when wi11 become apparent in this the European biospeleologists Rene chapter: Jeannel (France) and Candido 1. Western Mississi ppi an Plateau Bol ivar y Piel tai n (Spain) visited (NP-I)--Pennyroyal + Cumber- Mammoth Cave, Carter Caves, and land Saddle other caves in 1929 (Bolivar and 2. Eastern Mississippian Plateau Jeannel , 1931 . (MP-11)--Cumberland Plateau The bulk of our present know- Margi n ledge of Kentucky cave life is 3. Flue Grass based on widespread col 1ect ions 4. Karst Islands made by T. C. Barr, J. R. a) Pine Mountain Holsinger, T. 6. Marsh, R. M. b) Rough River Fault Zone and Norton, and S. B. Peck in the other isolated caves along 1960's. This material is still the outer margin of the being inventoried and described. Western Kentucky Coal Important contributions to cave Field taxonomy were made by J. M. Valentine (1952) who discovered the three endemic genera Darlinq- KINDS OF CAVERNICOLES toned, Ameroduvalius, and Nelsonitgs in the Somerset-Mount The taxonomic groups of animal s Vernon area, and by C. H. Krekeler in Kentucky caves will be consid- (1973), who collected Pseudanoph- ered in a subsequent section. thalmus species in the Blue First, we have to consider var-ious Grass. The richer and better known classifications of the ecological fauna of the Mammoth Cave Region or evolutionary status of caverni- attracted ecologists and other coles. Any animal that lives in a biologists who were dependent on cave is a cavernicole. If*we wish pre-existence of a taxonomic data to emphasize degree of restric- base. These biologists included T. tion to caves and other subterran- C. Farr, T. C. Kane, R. A. Kuehne, ean habitats (phreatic ground- T. McKinney, R. M. Norton, T. L. water, "microcaverns" around tree Poulson, and others. Some of the roots and in talus slopes, and so history of cave biology in the forth), we can divide caverni- United states was sketched by Barr coles into: SPECIAT'ION IN CAVES fi) troglobites--obligate cavern- SPECIATION IN CAVES icoles (e.g., blind cave- fishes) Speciation in cave faunas was B) troglophi les--f acul tative recently reviewed by Barr and cavernicoles that can live out Holsinger (in press). For Kentucky their entire lives in caves troglobites, the most common but are a1so found in non- pattern has been multiple cave microhabitats (springs, invasions of different cave deep soil) systems by widely distributed C) trogloxenes--caverni coles that non-cave ancestors; with gradual roost or remain in caves by extinction of the ancestral day but depend on feeding out- population outside caves, the cave side at night; "threshold" popul ations became isol ated and trogloxenes are bats, cave underwent genetic divergence to , and cave crickets varying degrees. The descendant D) accidental s--0ccasi onal cave species now occupy one to wanderers and that many caves, depending on (a) their wash or fall into caves mobility, (b) their capacity to Flatworms and small disperse through non-karst that disperse through groundwater terranes or not, and, if not, (c) in non-karst regions (and thus whether the caves they initial1y have wider geographic ranges than occupied are located in small, species unable to do this) are very local karst islands, or sometimes called phreatobites or broadly contiguous MP karst with stygobionts, which are more subterranean connections between general terms than "troglobite," caves, or some intermedite karst but they are, of course, area such as part of the Blue troglobites, too. Grass. Troglobi tes are re1ict species Even in MP areas there are that can no longer exist outside occasional dispersal barriers. In of caves or other subterranean MP-I these barriers are the Ohio microhabitats of the sort River, the Hart County Ridge and discussed; typically, their fault zone just north of non-cave ancestors are extinct in Munfordville, the Barren River at the cave areas, because surface Bowling Green, stratigraphic conditions hundreds of thousands barriers northwest of Hopkinsvi 11e or millions of years ago were and between the karst regions of different. Troglobites often show Caldwell and Crittenden counties, loss or rudimentation of eyes, and once again, the Ohio River. In pigment, circadian rhythms, and MP-I I the Mississippian 1imestone other characteristics associated outcrops are patchy and isolated with life above ground. The longer northeast of Red River, creating they remain in caves, the more several barriers. To the southwest successfully they have become there are river barriers, notably adapted to a cave environment, the deep gorge of Kentucky River showing such features as slender some 150 m below cave levels at bodies, long and stiltlike legs, Irvine, Cumberland River near longer antennae, hypertrophied Burnside, and South Fork as far as sensory systems, and so forth. the mouth of Little South Fork. Troglobites that have such River barriers rare1y af f ect adaptations are sometimes said to dispersal of aquatic troglobites, be high1y "troglomorphic, " a term but they effectively prevent that emphasizes evolutionary, dispersal of many terrestri a1 rather than ecolgical status. (For troglobites. Smaller, youthful further discussion of terms rivers with a meander frequency applied to cavernicles, see Barr, higher than l.O/km are rarely 1968a; Earr and Hol singer , in barriers, apparently because blind press). beetles and other troglobites can CAVE LIFE OF KENTUCKY be washed out of caves on one side different evolutionary rates and and manage to enter crevices in may eventual 1 y become four bi ol og- limestone bluffs on the opposite ical species (Christiansen and side (Earr, in press). Culver, 1968). Compared to tre- Di ff erent groups of trogl obi tes chine beetles, EL hirsutg has may show very small ranges--often either evolved more slow1 y or single caves--on the one hand, and entered caves more recently. quite large geographic ranges on Three species of blind , the other. Pseudoscorpions of the Orconectes eellucidus (MP-I genus Kleptochtonius, for example, south of the Hart County Ridge are rarely known from more than Fig. 1 Q inermis (MP-I north one or two caves, and distinct of the Hart County Ridge), and 0: species are reported from Mammoth ------australis (MP-I1 from Rockcastle Cave and nearby White Cave. County southwest to Alabama), are Smaller cave carabid bettles believed to be descendents of a (trechines) such as Pseudanoph= common epigean ancestor related to thalmus inexpectatus may be lim- 0: limosus, a non-cave species ited to Mammoth, White, and Great of the eastern At1 antic Coastai Onyx caves, but large, active Plain that lives in slowly moving, beetles such as Neaphaenop~ sluggish streams. In early Pleis- tell kampf i and. its close rela- tocene time, stream gradients tives in MP-I and Darlinqtonea steepened, and flow was faster, ------kentuckensis in MP-11 have very creating unsuitable habitats for 1arge ranges. the ancestral crayfishes, which The process of becoming a trog- retreated into small tributaries lobite (or whole series of troglo- and ultimately springs and caves. bitic species descended from a Hobbs and Barr (1972) postulated common ancestor that colonized survi val of thi s ancestral speci es different caves) seems to be in in three river systems--the Green, different stages at the present the Cumberland, and the preglacial time in different groups of Ken- Teays River: the three colonies tucky troglobites. For example, gave rise to eellucidus, 0, carabid bettles include about 60 ------australis, and & inermis, to 70 discrete species in Ken- respectively. These large troglo- tucky that probably became iso- bites are limited to the MP lated in caves by changing regions, for the most part climates during Pleistocene australis has penetrated the time. No ancestral species exist eastern edge of the Central Basin anywhere in eastern United States, in Tennessee in two places, one although one such colony evolving into the species (close1y related to isolated species 0: incogptus)A Greenbri er Val 1ey cave species because of numerous sol ut i onal nearby) is known from deep soi 1 in openings permitting dispersal. the Yew Mountains, Pocahontas Some "species" of troglobi tes County, West Virginia (Barr, are probably clusters of sibling 1967a) . Many deep soi 1 speci es of species--species difficult to trechine carabids are known in differentiate morphologically but mountains o+ Europe and eastern belonging to entire1 y different Asia. The widely distributed trog- gene pools. The likelihood that 1omorphi c col 1embol an Pseudo- IyehlLchthys sybterraneus a spe- sine/&g hirsutg (MP-I and MP-11, cies of blind whose range Kentucky and Tennessee) probably includes southern Kentucky (MP-I consists of populations descended and MP-I I ) , central Tennessee, from at least four separate cave northern Alabama, and the southern invasions in different parts of Ozark Plateau, is actually a its present day range. These four cluster of genetically differing groups have independently acquir- sibling species is quite high ed troglomorphic characters at (Barr and Holsinger, in press: 150 CAVE ECOLOGY

Figure 1. Orconectes ellucides (Tellkampf), the troglobitic crayfish of the Western Mississlpplan\ ateau.

Swofford and others, 1980). This pretation is correct, one might morphological species probably now anticipate that the end result consists of surviving relict pop- wi11 be a large number of local, ulations of a wide1 y distributed mare discrete, and morpholog- ancestor that has become extinct: ical1 y more distict species. troglomorphic features have been acquired, as in Pseudosi nel la CAVE ECOLOGY hirsuta, by para1lel evolution. Phanetta , a tiny Physical aspects of the cave troglobitic spider originally environment have been f air1y we1 1 described from Carter Caves, described (for Kentucky, see Earr Kentucky, is perhaps the most and Kuehne, 1971): the deep cave widely distributed troglobite in is characterized by absence of the eastern United States. Even if 1i ght , constant temperature equi v- it is capable of dispersing read- alent to the regional mean surface ily through microcaverns in non- temperature, low vapor pressure karst terranes, it is inconceiv- deficit, slightly alkaline pH of able that gene flaw is maintained cave waters, and so forth. Season- throughaut this whole complex od al fluctuations, however, may be populations. Most likely. PL profound. Cold, dry, winter air subterranea is a relatively flowing into entrances expands to recent cave colonizer, and it has cave temperatures and soaks up been extremely successful in col- moisture from the walls and spele- onizing a large number of caves othems, so that the rate of evap- over what must have been the oration may be as much as 200 range of its extinct, non-cave times that of the same part of the ancestor. In time, if this inter- cave in summer. Most terrestrial CAVE LIFE OF KENTUCKY troglobi tes cannot tolerate low pseudoscorpions, carabid beetles). relative humidities and disappear Two groups of carabid beetles, from entrance zones in winter. Neaphaenops in MP-I and Darlinw Aquatic cave microhabitats are -----tonea in MP-11, have indepen- subject to seasonal differences in dently "learned" to dig and eat water levels in pools or to heavy eggs of these two flooding in streams. species (Hubbell and Norton, 1978). Hadenoecus cumberlandicgs Ultimately, all food utilized has evolved parthenogenetic popu- by a cave ecosystem comes from lations (females only) in WP-I1 the surface. There are two ways northeast of the in which food gets into caves: (1) (Lamb and Willey, 1975: see also transport by vadose water, and (2) Glesener and Ti 1man, 1978) . transport or direct contribution (feces, dead bodies) by troglox- Larger streams sinking under- enes such as bats and cave crick- ground may carry leaves and twigs ets. Cave rats (Neotoma maqister) that are deposited along cave contributes leaves, twigs, and stream banks and eventually decom- feces to the deep cave community; pose through bacterial and fungal such food is usual1y limited to action. Vadose water entering shelves or the. floor below shelf caves vertically along joints and runways or nest sites. Hiberna- fhrough dome pits may leach out ting bats contribute very little, organic compounds from humus in but summer bat residents contrib- the overlying soil mantle, at ute large amounts of guano. Only the same time carrying with it the gray bat, Myotis grisescens, soil bacteria, protozoans, and is a significant contributor of other microorganisms. Vadose water guano in Kentucky caves, princi- samples in Mammoth Cave measured pally in the southern part of the at different seasons of the year Commonwealth. Bat guano decomposes have a total organic carbon con- slow1y, often with toxic byprod- tent ranging from 2 to 14 ppm ucts, and seems more significant (Barr, unpublished). These organic as food when it is wide1 y dis- compounds-hexoses, amino acids, persed on a wet floor, under flight humic and fulvic acids, and so paths, or when it falls in small forth--are apparently adsorbed quantities into rimstone pools, onto clay micceles in the mud of than when it accumuates in large cave pools. Small troglobitic mounds. Far more important for crustaceans--amphipods and Kentucky cave communities are cave isopods--that eat their way crickets of the genus Hadenoecus, through the mud in pools can H, subterraneus in MP-I and Hi probably release the organic com- gggberlandicus in MP-11. Unlike pounds by altering the pH in their the common cellar and camel crick- digestive tracts. 0 wide range of ets (Ceuthophilus spp. ) , which soil microorganisms enter caves in rarely penetrate deeply into dripping water, but relative1y few caves, these two species (along survive for more than a few weeks. with thr,ee simi 1 ar species in Most cave protozoans, for ex amp1 e , MP-I I of Tennessee and A1 abama) belong to a small number of roost in large numbers in favored ciliate species that are repeat- sites on cave ceilings, emerging edly encountered in various cave at intervals of 2 or 3 days to waters (Gittleson, 1969). Although feed outside. Beneath the roosts a there are no photosynthetic pro- thin, humus-like layer of guano ducers in cave communities, accumulates, providing food for bacteria and fungi, through decom- many detritus-feeding troglobites positon of organic materi a1 washed (snails, col lembolans, bristle- into caves, act as secondary pro- tails, Ptomaphaq~~sbeetles, etc. ) . ducers, transforming the material The detritivores are in turn eaten into a form that can be utilized by predatoy troglobites (, by troglobites. In the "Shrimp CAVE ECOLOGY

Pools" in the Roaring River mud of stream banks. These worms, Passage in Mammoth Cave, bacter- in turn, provide food for preda- ial counts on a wide spectum agar tory carabid beetles, many species are lowest in spring, following of which occur--particular1 y in annual floods. As the summer wears the western part of MP-I--in great on, counts rise and coninue to abundance along the banks of cave rise, reaching their maximum just streams. A few species of carabid prior to the first flood of wlntrr beetles feed principal 1y in upper (Farr and Kuehne, 1971). The levels of caves, where they occur Mammoth Cave blind shrimp. in the wet, rotting debris of old Palaemonias ganteril (Fig. 2) cave rat , picking thrnuuh feeds on microorganism5 by strain- the debris and eating the small ing bottom muds through their detritivores found mouthparts. there.

Bacterial and fungal decompo-=.I ' - Several striking ecological tion of stream-borne detritus pro- differences have been noted be- vides food that can be utilized by tween caves of the MP regions and small, threadlike segmented worms caves of the Appalachian Val ley (enchytraeids and tubificids, A (Parr, 1967b). Extensive possibly undescribed troglobitic folding and faulting has produced species) that through the patchy, quite local karst areas in

Figure 2. Palaemonias ganteri Hay, a troglobitic shrimp known only from the Mammoth Cave system. CAVE LIFE OF KENTUCKY the AV. Although the AV does not must be mutually compatible, not extend into Kentucky, similarly competing for the same environ- local and stratigraphicall y iso- mental resources. Several ser ies lated karst islands in Pine of close1y similar species of cave Mountain, the Rough River Fault carabid beetles are known that are Zone, and patchy exposures of Glen parapatric, ti. e. , their ranges Dean Limestone around the edge of are contiguous but non-over- the Western Kentucky Coal Field lapping). Following colonization are ecologically and biogeograph- of two or more different cave ically similar to AV caves. In the systems by the same ancestor, MP regions the dispersal potential genetic divergence has occurred. of a troglobite is much greater: As the different beetle popula- it can move from cave to cave by tions dispersed outward from their subterranean solutional openings points of origin by subterranean in broadly contiguous karst. In routes, they eventually came into contrast to AV or karst island contact with each other. Because caves, troglobite species in MP of gnetic differences, hybrids caves have larger geographic between the two populations were ranges, exhibit fewer species per unit area of exposed karst, and inferior; consequently, natural are more commonly sympatric (i.e., selection favored in-group breed- ing. Ecological niches of the two two or more species coexist in the populatons had not diverged enough same caves). 63 typical troglobitic so that they could coexist in the communitv in the MP regions< may include 15 to 30 species, while same caves. As a result of genetic only 5 to 10 species are found in incompatibility and niche similar- AV or karst island caves. MP trog- ity, the two populations estab- lobites are often larger than lished a range boundary that those of AV or karst island caves, neither could cross. Much of the a phenomenon possibly attribut- western part of MP-I is inhabited able--at least for predators--to a by species of the pubescens group wider variety of prey species, of Pseudanoghthal mus: most show consequently a more predicatable parapatric ranges. The sole excep- supply of food. In a small, iso- tion is a distinctly smaller lated cave system there may be species with different feeding only one or two prey species, and habits, PI hganensis (Barr, a carabid beetle predator depen- 1979). While most species of the dent on these will suffer a de- pubescens group are larger and cline in number if the prey cruise mud banks in search of species decline. In MP caves, prey, PI Ipqanensis hides in factors that affect one prey detritus piles and between clay species may not necessarily affect laminae, feeding on small arthro- all of them, consequent1 y carabid pods that it finds there. In beetle populations are often not various caves P, loqanensis co- only unusually large but much more exists with PL ciliaris and PI stable (Barr, 1967b, 1968, in prince~s~both larger species of press). the pubescens group. The mean The opportuni ties for evolution 1engths of most 1oqanensi s popula- of complex communities of troglo- tions are close to 4.5 mm, while bites are much more favorable in ciliaris is normally close to 5.0 MP areas. Different ancestars may mm mean length. Perhaps it is sig- have colonized different caves at nificant that all ciliaris co- different times in the past, but existing wih loqanensi s belong to because there are extensive the distinct1y larger subspecies connections between MP caves, sub- PI c, prggndae, with a mean terranean dispersal may permit length close to 6.0 mm, thus ex- many species to eventually occupy acerbating the size difference the same caves. Coexisting species (Earr, 1979, and in press). 154 CAVE ECOLOGY

The evolution of cave communi- species with different environmen- ties in the Kentucky karst appears tal requirements. Furthermore, it to be conditioned by three circum- lies at the edge of the Pennyroyal stances: (1) which groups were Plateau, a major avenue of troglo- present to colonize caves in the bite dispersal because of its net- first place, an historical factor; work of interconnected caves: con- (2) the extent of contiguous sequently, its fauna is composed karst, which affords troglobites of species with both northern and the opportunity to eventually southern origins. In addition, the reach the same caves, and (3) Mammoth Cave community has ele- niche compatibility of component ments that have perhaps dispersed species. By the principle of com- into the area from the southeast, petitive exclusion, two species across the Cumberland Saddle with essential1y identical niches (Barr, 1968b). More trechine cara- cannot coexist in the same cave bid beetles occur in Mammoth Cave community. than in any other North American The Mammoth Cave community has cave-- Neaphaenops tellkampfl been studied more often than any (Fig. 3), and five species of other cave community in North Pseudanophthalmus. The 1argest America. filthough more than 200 species, N, tellkampf i, feeds species have been recorded from heavily on eggs of Hadenoecus the system, only about 30 are com------subterraneus, and its own repro- ponents of the deep cave, troglo- ductive cycle usually tracks that bitic community. As an extremely of the cave crickets, supposedly large cave system with many because avai 1abi 1i ty of abundant different microhabitats, Mammoth food (cricket eggs) permits egg Cave can accommodate a variety oi production in the beetles (Norton

Figure 3. Nea haeno s tellkampfi (Erichson), a 7 mm trechine beetle from the Mammot--Eh---E Cave area. CAVE LIFE OF KENTUCKY and others, 1975). Two smaller Protozoans and Other Microfauna species, P, inexpectatus and PI Host protozoans found in caves audax, appear in wet weather and are small ci1 iates (Gittleson, unusually dry weather, respec- 1969). They are morphologically tively, suggesting restriction to simi lar to non-cave species. quite different niches (Farr, Nematodes, harpacticoid copepods. 1966-67). The flattened form of and creeping rotifers occur in cave e, gudax and its great rarity in pools and streams but have not been caves suggests that its normal studied in detail. microhabitat could be microcaverns at the interface between bedrock Flatworms and soil mantle. The three other Most cave flatworms belong to species are subequal in size the genus Sphallmplana, whose spe- (about 5.5mm) -- P, mnnetriesi , cies are widely distributed in ztriatus, and pubescens. PI PI P, Kentucky (Kent, 1977). Geocentro- Both menetriesi and striatus be- phora cavernicola, however, is long to the menetriesi group, and known only from a single MP-I1 thus have higher taxonomic af f in- cave (Carpenter, 1970). ity to each other than either has to pubescens, which belong to the Snails pybgscens group (Farr, 1979). Carychium styqigm is a minute, While striatus is almost ex- P, P, troglobitic land snail only 2 mm clusivel y a riparian species, high (Fig. 4); it feeds princi- found along stream banks in the pally on cave-cricket guano and is lower levels of caves, men- P, limited to MP-I from Hart County g&riesi 's normal habitat is piles to the Tennessee line (Hubricht, of debris in upper levels. Pseuda- 1950) . Several other snai 1s from balmgz pubescens, 1 ke cpphf i el Kentucky caves were described by striatus, is a riparian feeder, ------Hubricht (1962, 1963, 1964, 1965, although it can appear in consi- 1956, 1968a. l9bBb); their status derable numbers in upper levels --troglobites or not--is uncer- during periods of unusually wet tain, but some appear to be limit- weather, suggesting that its ed to caves. Antros~lates normal microhabitat has been spiralis is an aquatic troglobite f 1ooded out (Barr 1966-67: , known from Echo River in Mammoth McKinney, 1975). Cave (Hubricht, 1963). Cave communi t ies in karst islands of the Western Kentucky Pseudoscorpions Coal Field or Pine Mountain are About a dozen troglobi tic much less complex, comprising on1y pseudoscorpions are known from to 10 troglobitic species. Cave 5 Kentucky caves, but there are communities of the Blue Grass are undoubtedly many more, especially more or less intermediate in eco- in the genus Mleptoch~honius logical characteristics between the (Muchmore, 1963. 19651, most two extremes of the MP and AV, species of which appear restricted having a rather small number of to one or two caves (Fig. 5). component troglobite species, but some of these species have geo- Harvestmen graphic ranges that are moderately Phalanqodes armata is an eye- extensive, depending on contiguity less, pale harvestman, or opi- of karst. lionid found in the Mammoth Cave and Bowling Green areas (Fig. 6). A BRIEF SUMMARY OF MAJOR Troglobitic species of this group GROUPS OF CAVERNICOLES are very limited in Kentucky: the status of Erebornaster flavescens Emphasis in this section is on caecum in Carter Caves is uncer- more conspicuous faunal groups, tain, but it is usually thought of especially troglobites. as a troglophile race of a more 156 SPIDERS

Spiders The most abundant spider in Kentucky caves is Meta menardis the cave orb weaver: it is normai- ly a threshold trogloxene. Four widely distributed morphospeci es of tiny cave spiders in the fam- ily Linyphiidae occur in Kentucky -- Phanetta s~tbtercgneg(type locality, Carter Caves. Kentucky), !?ochornrna-~avicolurn, Bathyehanz fez ~eyeri~and Anthrobla ~outhia(type locality. Mammoth Cave, but also found in Tennessee. Virginia, and West Virginia). Nesticus carteri is common in MP-I1 caves: it is a troglophile. A handful of other spiders are commonly found in the twilight zone: they are not troglobites (Barr, 1968b). Mites Little is known about cave mites. Hhaqidia cavernarum, f rom Mammoth Cave, is rather widely distributed (Holsinger, 1965). Several other genera and f ami 1 ies are represented in caves but appear to occur erratically and may not be troglobites. Amphipods Two groups of troglobitic (styqobiont) amphipods occur in Kentucky caves. The species of S3yqobrorn~ts are bel ieved to be much older in terms of residence in subterranean waters: they occur in quiet pools, usually in food- poor microhabitats. Cranqonyx packgrdi, on the other hand. is the central Kentucky (MP-I to MP-11) representative of 5 cave species that are relative1y recent cave invaders: it is most likely Figure 4. to occur in food-rich streams and Call. from pools (Holsinger, 1978, in press). retains vestigial eyespots but ' is a troglobite that feeds on lsopods cave cricket guano. Troglobitic (stygobiont) iso- pods of the genus Caecidotea are common in a majority of Kentucky widely distributed Ohio Val 1 ey caves. These include species, such species. The 1arge grayi sh-brown as C, stygig in Mammoth Cve, with harvestmen that gregariously very wide distributions, but also hibernate in caves are usually species that are rare and quite Leiobgngm lonqipesL local in distribution. The group CAVE LIFE OF KENTUCKY

Figure 5. Kleptochthonius sp. is typical of many troglobitic species in this genus that are seldom found in more than one or two caves each. was recently reviewed by Lewis (Earr 1968a, 1968b; Hobbs and (1984). others, 1977). Crayfishes Centipedes &s noted in the section on No troglobi tic centipedes are "Speciation" above, three trog- known in Kentucky caves, and the lobitic species of Orconectes occasi onal species that appear seem occur in Kentucky, 0, pex?&ucid~~s to be accidentals. and Lnermis in MF-I, and QL Millipedes -austral------is- in MP-I1 CHobbs and Barr, 1972). Cambarus tenebrosus In contrast to centipedes, the is a tr~glophilecrayfish that millipedes are quite well repre- occurs frequently in most cave sented. Species of F'se~tdotremia regions in Kentucky; it often are found in many caves; some of attains unusually large size. them are very pale troglobites. whi le others--more wide1y distri- Shrimps buted--ars troglophi les. Six are Palaemonias qanteri%restrict- recorded from C::entucky, but there ed to river levels in the Mammoth are undoubted1y more (Shear, Cave system, is an isolated relict 1972). Scoterpes occur in many species; its closest relative caves in the southern MP-I and lives in Shelta Cave, Alabama MF-I I; a1 1 of its species are 158 COLLEMBOLANS

Figure 6. Phalan odes armata Tellkampf, the troglobitic harvestman of the M~.mmoth&owlingGreen areas, is a predator. troglobites. The genus has not ferent family, . been systematically studied, and Species of Ceuthophilus are shiny, the Mammoth Cave area SL copei is banded cave crickets (in Kentucky) the only one of several Kentucky that are often abundant near en- species that has been described. trances. (MP-I ) and HI cumber1 andi cgs Collernbolans (MP-11) are ecologically more im- Several "springtails" inhabit portant because they are restrict- Kentucky caves, includ'ing the ed to regions with numerous caves widely distributed morphospecies and deposit much guano beneath Pseudosinella hirsuta, the highly their roosting sites. Both are troglomorphic P, christianseni trogloxenes, not shiny, not and Sinella krekeleri CMP-11) , banded, extremely 1ong-l egged, and other troglobitic and troglo- with long antennae. They are philic species (Christiansen, absent from the Cumberland Saddle 1960a, 1960b, 1966: Christiansen region of MF-I (Hubbell and and Culver, 1968). Nortan, 1978). Diplurans Trechine Beetles These si1 verf ish-1 ike "bristle- About 65 species of these tails" belong to the genus !Atpr small. predatory carabids occur In campa and have been recently Kentucky. Most belong to Pseuda- studied by Ferguson (1981). Ranges nophthaim~~s,but there are 1 to of the species recognized by him 5 species each in ijeaphaenopz are re1ati vel y estensi ve. suggest- tMP-I). Nelsonites (MP-119. Day: ing either microcavern dispersal -1 -i nqtonea------iMF-I I), and Amer- or groups of close1y similar bio- pcigyal iES !MF-I I ) . The re1atea logical species (Fig. 7). troglophi 1e Lrechus cumber1 and~ts also occurs in caves and sinkholes Cave Crickets in MF-11, though it is not ances- These are not true crickets, tral to the cave species. Eleven instead belong to an entirely dif- of the 26 species groups of CAVE LIFE OF KENTUCKY

Figure 7. Litocarr a cookei (Paclcard) , widely distributed in Kentucky caves, is a +etrltus feeder eaten by beetles, pseudo-scorpions, and other predatory troglobites.

Pseudanoehfhalmus, which has in Tennessee) (Park. 1960, 1965). about 240 species in 10 eastern ~t~rnaehas~z-l~delnesLhIr_t~rz states, are represented in Ken- (Leiodidae) is a detritus-+eedinq tucky, and both Neaphaenops and troglobite in the Mammoth Cave Ameroduvalius are limited to Ken- area, where it may be locally tucky. Descriptions of most of the quite abundant: another 17 troglo- k:entucky species are in press or bitic Ptomaphaqgs spp. occur far- in preparaton (Valentine. 1952; ther south in Tennessee and Barr, 1?7?a, 1979b. 1981, 1985, in Alabama (Peck 1973, 1984). press). Fishes Other Beetles A variety of fish species may Pselaphi d beet1es are small be washed accidentally into caves. predators that superficially and sculpins actively swim up- resemble because of their stream into caves, but the only shortened elytra. Kentucky cave troglobitic species in Kentucky pselaphids are less troglomorphic belong to the Amblyopsidae. than some genera and species in hblyopsis seelaea, described Tennessee and Alabama: they in- from Mammoth Cave in 1842, clude a few species of Batrisodes ranges northward in MP-I into 2 species of Sjythinopsis, and the southern Indiana. Iyphli chthys wide1 y distributed troglophile subterraneus a1 so occurs in Batriasymmodes quisnamus (south- Mammoth Cave, but nowhere else central Kentucky and more common does its range overlap that of SALAMANDERS

Ambl ygesis spel aea. In Ken- Eetesicus fu~cus~are more soli- tucky it is reported from MP-I tary bats, rarely occurring in south and west of Mammoth Cave, large hibernating groups. Pipis- including caves in Glasgow and trelles are perhaps the most Bowling Green and from the Sloans common bat in Kentucky caves, but Valley System in MP-I1 !Cooper !EL f~tsc~ishibernates in colder and Eeiter, 1972). Choloqaster sites, often near entrances in the -aqassiri ------is a troglophilic ambby- fall and early winter. Both the opsid found in caves and springs eastern and western species oi of southwestzrn MP-I and elsewhere long-eared bats are found (Woods and Inger, 1957; Poulson, lusually uncommonly) in Kentucky 1963). caves: Plecotus rajlnesquel f rom the Mammoth Cave Region and Salamanders Jackson, Pulaski , and Wayne No troglobi tic salamanders counties; and E, townsendi occur in Kentucky, but the "cave virqlnianus from a small number salamander " Eurycea 1uci f uqa is of caves in Lee County (MP-II), widely distributed; it is more where it is in need of protection. common in entrance areas where food is more readily available. Eseudstritnn ruber, Ey~yr~a REGIONAL CAVE FAUNAS IN KENTUCKY ---lonqicauda, and Plethodon dor- salis are reported from occa- Western Mississippian Plateau sional caves in MP-I and also in (MP-1) and Cumberland Saddle the Rough River region. The Mammoth Cave faunas is one Cave Rats of the richest in the world (Barr, 1968b) , including troglobitic Neutoma mtggister is the common species of f 1atworms, snai 15, "cave rat" of Kentucky caves, al- pseudoscorpions, harvestmen, though floridanus, common in spiders, mites, amphipods, isopods, Tennessee, may occur along the crayfishes, shrimps, millipedes, State border. col lembolans, diplurans, beetles, and cave fishes. A majority of the Bats species in Mammoth Cave itself Ecologically, the most impor- range northward toward Wunfordville tant bat species for Kentucky and southwest to Barren River along troglobites is the gray bat, the Pennyroyal, but to the Nyotis qrisescens, which lives southeast, across the Cumber1 and in caves in summer and is thus a Saddle, the composition of the guano bat. Most of its colonies fauna changes rapidly--most beetle are in the southern part of MP-I. species disappear or are replaced and there is evidence that many by others; and crayfishes, older colonies have been aban- cavef i shes, and even Hadenoec~ts doned. The largest populations of subterrane~tsdi sappear. hibernating bats +re Myotis To the north, few terrestrial soda1 15%the "social, " or Indiana species cross the Hart County bat; this species migrates to Ridge and +ault zone (Barr 1968b, northern, non-cave roosting sites 1977a, in press). A somewhat in summer: it has been extensively diverse fauna continues on to the studied in the Carter Caves and Ohio River, characterized by Mammoth Cave areas as well as a !lrcnsectes inecmis, B~blyqesis few other widely scattered caves, seelaea (8arr and Kuehne, 1962), mostly in MP-I and MP-11. Smaller a few species of Pseudotremi a, colonies of the little brown bat, Negphaenops henroti, and Pseud- tlyotis l~tci+uqus, hibernate in a anoehthal mus of dif f erent speci es number of Kentucky caves. The than occur south of the Hart pipistrel le, Pipistrellus sub= County barrier. Hadenoecus sub- flavus, and the big brown bat, terraneus is abundant. CAVE LIFE OF KENTUCKY 161

To the southwest, across Barren A small fauna (2 species of River, Lyphl ichthys s~tbtg-raneus Pseudan~ehthalmuz, Sin~lla, is present but not fimblyopsis Eleetochthnnius, Pseud~t~emia, spelaea; all trechine beetle Crangmyx, Caeciodotea, bisexual species are different, and even Hi cumber 1andi cus, etc. ) occu- HI subterraneus varies geo- pies upland caves between Red and graphical 1y. Orconectes Kentucky rivers in Powell, Estill, pel1 ~tcidus is present. and the edge of Lee counties. The Farther west in the Pennyroyal. Kentucky gorge forms a major bar- Neaphaenops disappears at the rier to terrestrial troglobites; eastern edge of Logan County, and south of the Kentucky River, three Hatienoecus is no longer present endemic genera of trechine beetles west of Russel 1vi 11 e. Scoterpes appear-- Darlinqtonea, Amerodu- copei is replaced by one or more ------val ius, and Nelsoni t es . undescribed species oi the same Scnferees ~PO,Occqnectes genus, and the prevalent trechine ------austral is packardi (from beetles are wide-ranging, poly- Rockcastle County southeastward), typic species (such as P, and Iyphlichthys (in Sloan Valley ciliaris; (see Barr, 1979a) whose System) augment the fauna farther ranges do not overlap. Both 0, to the southwest. F'se~tdanoeh- pellucidus and 1, subterraneus thalmus species in the range a+ continue at least to Trigg County. Ameroduval ius species are a1 1 Distinct trechine species with small, less than 4 mm long, but discrete ranges occupy the - south of the Cumberland River Frinceton-Freoni a and Sal em-Mar ion there are several larger Pseuda- cave regions, respectively. n~phthalmus of the robustus Southeast across the Cumberland group. Wayne County, Kentucky, Saddle from Mammoth Cave there is with 12 species and 5 genera of a moderate fauna of trechine trechine beetles, has the most beetles, some millipedes, amphi- diverse trechine fauna of any pods and isopods, but Orconectes, comparable area in North America. Neaehaenoes, Hadenpecuz, and There is a partial dispersal bar- Lyphlichthyg are not present. The rier near the Kentucky-Tennessee upper reef limestone member of the border in MP-11, because westward contains a flowing tributaries of the number of caves that support a Cumberland do not join until they northward extension of this dauna have cut down into non-caverni - to the Columbia and Greensburg f erous Osagian rocks. The area. The eastern extent of Cumberland River near Burnside NeaE$aenoes, Hadennecus, and constitutes a triple barrier near Orcon~ctesis the vicinity of Burnside, Pulaski County: the Greensburg, itself. lower Cumberland River below the mouth of South Fork is a complete Eastern Mississippian Plateau barrier for terrestrial beetles, as is the lower South Fork below North of Red River the Newman the mouth of Little South Fork. Limestone thins and is confined to The upper Cumberland acts as local lenses thick enough to though it were a partial barrier, support caves; much of this area permitting but reducing gene flow: has only dry, rockhouse-like rem- the upper South Fork is no barrier nants of former caves on top of at a1 1. the ridqes. However, local faunas including some troglobites are known for the Carter Caves area in Blue Grass Carter and El1 iott counties, Caves of the Inner Blue Grass Murder Cave in Menifee County, and are developed in more or 1ess con- one or two additional such karst tinuous patches of Ordovician isl ands. Hadenoecus cumber1 and- limestones. Various species of ----icus is all parthenogenetic here. spiders, isopods, and millipedes KARST ISLANDS are known from single cave sys- Pseudotr~mia, Pseudosinel 1a tems, but some of the Pseudanoph- hirsyta, Litocamea, Caeciz &ha1mus species are moderate1y dnfea, St~snbromvs~and Sehalnz widespread, notably between plana are also found in Pine Clif ton and Camp Nelson a1onq the Mountain caves. right bank of the Kentucky River. Karst isl ands a1so occur around A population of parthenogenetic the margin of the western Kentucky !-jadenoecus cumberlgndicus is Coal Field near the edge of MP-I: known in small caves around Camp almost all significant caves in Nelson: it is geographiclly iso- these areas are developed in Glen 1ated from bisexual popul ations of Dean Limestone, stratigraphically the species to the east. Trechines isolated from the thicker and more include species of the inexeec- cavernous limestones of the Penny- tat~tsgroup, related to P, royal. Most of the troglobitic ----inexpectatus ------in Mammoth Cave, and species are wide-ranging ones with a1 so species of the hprni group, means of dispersal through mirro- which is .distributed not along caverns or phreatic water, but at present drainage basins but a1ong least 9 isolated species of the preglacial Teays Val ley and i>zeudanophthalmgg are known (five includes species in southwestern from the Rough River Fault Zone in Ohio and southeastern Indiana. Grayson County: others in Hart, These beetles are smaller (3-4 m) Warren, Butler, and Todd coun- than those of the MP region. ties). The beetle species are Pseudanophthalmus bgrri near closely similar to species of the Clarksvi 1le. Indiana, and EL Pennyroyal surface caves and are fygqlphytes are sister species presumably derived from the same that were probably separated by common ancestors. Hadenoecus development of the modern Ohio subterraneus occurs in all of River during early Kan'san time these marginal karst island (Krekeler, 1973). In MP-I farther caves. A small cave in Madison- west, ez tenuis (Indiana) and P, vi11 e Limestone (Pennsylvanian) barberi (Kentucky! bear the same near Equality, Ohio County, con- relationship to each other tains Sphal loplana, Cranqonyx (Jeannel, 1949). packardi, Sinellal and Phanetta, as well as numerous Karst Islands troglophi les and trogloxenes. Small, isl and-1 ike patches of karst occur on the northwestern CONSERVATION OF CAVE LIFE face of Pine Mountain in south- Much attention has been devoted eastern Kentucky and also downdip to conservation and preservation from MP-I, toward the center of of bat species, with emphasis on the Western Kentucky Coal Field. avoiding disturbance during hiber- These patches are characterized by nation. Growth of urban areas in small - assembl ages of isol ated spe- karst regions, with accompanying cies. For example, the endemic closure of caves and pollution of hueolithos group of Pzeeda~eeh= underground streams, notably at ------thalmus is represented by species Somerset and Bowling Green, poses in Pike, Harlan. Bell, and Whitley a threat to fish and invertebrate counties, Kentucky, and a fifth troglobites. Leakage from oil species just south of the A1 le- wells and pumping out wells with gheny Front in Virginia. One spe- brine has polluted many Kentucky cies of the ipnesi group is des- caves in rural areas. The evolu- cribed from a Harlan County cave: tion of Hidden River Cave into a the remainder of the group occurs gigantic sewer for the town of in southwestern Virginia and in Horse Cave or the construction of Pine Mountain (Campbell County, a garbage 1andf i 11 above Sl oans Tennessee), and Grassy Cove Valley Cave are tragedies that (Cumberland County, Tennessee). should never have been allowed. CAVE LIFE OF KENTUCKY

The larger troglobites of cave to major hibernacula caves at streams-- Oroconectes crayfishes critical times of the year. and the amblyopsid cavefishes-- Finally, permanent closure of face a long-term danger from even caves, either by bulldozers in an very low levels of heavy metals. urban area or by landowners who These species grow extremely slow- have been pestered by inconsid- ly and may require 25 to 50 years erate cavers, may or may not mean to attain sexual maturity. When extinction for the fauna of these troglobitic and non-troglobitic caves. Closure does mean that the crayfishes from the same cave in fauna of those caves is no longer Tennessee were analyzed for heavy available for study, by anybody. metal concentrations in their tis- Krekeler (1973) described tw~ sues, the troglobitic species were unusual species of cave beet1es shown to have hundreds of times on opposite sides of th Ohio the concentration of several River near Loui svi 11e, Fseuda- metals that were found in the non- nophthalmus barri in Indiana troglobitic crayfishes (which and PL t~oqlodytesin Kentucky. reach sexual maturity in about 2 The only cave where f, troq- years). Similar very high levels ---lodytes --- has been collected is were found in Qg~onectes now under Osmoor Shopping Center. pellucidus from Parkers Cave, Barren County, Kentucky. These REFERENCES CITED species live so long that minute concentrations of heavy metals Barr, T. C., Jr., 1966, Evolution acquired each year may ultimately of cave biology in the United reach toxic levels. Consider the States: 1822-1965: National Spe- cavef ishes and crayfishes of Echo leological Society, v. 28, p. River in Mammoth Cave, f ortunatel y 15-21. spared the Hidden River pollution. What about the minute amounts of Earr, T. C. , Jr. , 1966-67, Cave cadmium generated from wear of Carabi dae (Coleoptera) of Mam- vulcanized tires and of lead from moth Cave: Psyche, v. 73, p. 1eaded gas01 ine combusi on a1ong 284-287: v. 74, p. 24-26. Interstate 65? There is a very real possibility that the world- Parr, T. C., Jr., 1967a, A new famous fauna of Echo River could Pseudanophthalmus from an epige- become extinct, not from a disas- an environment in West Virginia trous pollution accident, but (Coleoptera: Carabidae): Psyche, slowly and insidiously from low v. 74, p. 166-172. levels of toxic metals. The fauna would disappear not with a bang, Earr, T. C., Jr., 1967b, Observa- but with a whimper. This is a tions on the ecology of caves: potentia1 danger that is seriously American Naturalist, v. 101, p. in need of further study. 475-492. There is not much hard evidence that heavy caving pressure, other Earr, T. C., Jr., 1968a, Cave than the dumping of spent carbide, ecology and the evolution of which is fortunately becoming a troglob?tes: Evolutionary Eiol- 'thing of the past, seriously re- ogy, v. 2, p. 35-102. duces popul ation 1eve1 s of trog- lobites. Disturbance of hiberna- Earr, T. C., Jr., 1968b, Ecolog- ting bats during winter or dis- ical studies in the Mammoth Cave turbance of maternity colonies of System of Kentucky: I. The bats are probably the greatest biota: International Journal of damage cavers are likely to Speleology, v. 3, p. 147-283. inflict on Kentucky cave life, but this can be avoided by education, Barr, T. C., Jr., 1979a. The tax- gating, and restriction of entry onomy, distribution, and affini- 164 REFERENCES CITED

ties of hlegphgengps, with notes United States: Psyche 67, p. 1- on associated species of 25. Pseudanophthalmus (Coleoptera: Carabidae): American Museum No- Christiansen, K. , 1960b, The genus vitates, v. 2862, 20 p. Sinella Brook (Collembola: Ento- mobryidae) in Nearctic caves. Barr, T. C. , Jr. , 1979b, Revision Anna1 s of the Entomol ogi cal So- of Appal achi an Trechus (Coleop- ciety of America, 53, p. 438- tera: Carabidae) : Brimleyana, 491. no. 2, p. 29-75. Christiansen, K., 1966, The genus Barr, T. C. , Jr., 1985, New Arrhopalites in the United trechine beetles from the Appa- States and Canada: International 1achian region (Coleoptera: Journal of Speleology, 2, p. 43- Carabidae): Brimleyana. no. 11. 73, pls. 11-14.

Barr, T. C., Jr., in press, Chri stiansen, K. , and Cul ver , D. Pattern and process in specia- C., 1968, Geographical variation tion of trechine beetles in and evolution in Pseudosinella eastern North America (Coleop- hirsuta: Evolution, v. 22, p. tera: Carabidae: ) , in 237-255. Pall, G. E., ed. Taxonomy, phy- logeny, and zoogeography of Cooper, J. E., and Peiter, D. P., beetles and ants: Series Ento- 1972, The southern cavefish, mologia, v. 33, p. 350-407. Tvehlichthy~ subterraneys (Pis- ces: Amblyopsidae) in the east- Parr, T. C. , Jr. , and Holsinger, ern Mississippian Plateau of J. R., in press, Speciation in Kentucky: Copeai, v. 4, p. 879- cave faunas: Ecology and System- 881. ati cs , Annual Review. Davis, W. H., and Parbour, R. W., Parr., T. C., Jr., and Kuehne, R. 1965, The use of vision in A., 1962, The cavefish, Ambly= flight by the bat fJypfL5 ppsis spelaea, in northern Ken- soda1 is: American Midlands Nat- tucky: Copeia, v. 3, p. 662. uralist 74, p. 497-499.

Earr, T. C., Jr., 1971, Ecological DeKay, J. E., 1842, Description studies in the Mammoth Cave sys- of firnblyoesis seelaea, in tem of Kentucky. 11: The Ecosys- Zoology of New York, or the New tem: Annals of Speleology, v. York Fauna, Part IV, Fishes: 26, p. 47-96. Albany, New York, footnote, p. 187. Pol ivar , C. , and Jeanne1 , Rene, 1931, Campagne speologique dans Ferguson, L. M., 1981, System- 1'Amerique du Nord en 1928 atics, evolution, and zoogeogra- (premiere serie) : Arch. zoo1 . phy of the cavernicolous campo- et gen. 71 , p . 383-388. deids of the genus Litocampa (: ) in the Carpenter, Jerry H., 1970, Geo- United States: Placksburg, Vir- centrophora cav~rnicola~n. sp. ginia Polytechnical Institute Turbel laria, A1 loeocoel a) : First and State Univiversity, Ph.D. cave a1 1oeocoel : American Micro- Dissertation, 372 p. scopical Society Transactions 89, p. 124-133. Gittleson, S. M., 1969, Cavern- icol ous Protozoa: review of the Christiansen, K., 1960a, The genus literature and new studies in Pseudosinel 1a (Collembol a, Ento- Mammoth Cave, Kentucky: Annals mobryidae) in caves of the of Speleology 24, p. 737-776. CAVE LIFE OF KENTUCKY 165

Glesener, R. R. , and Ti lman, D. , podinae) . University of Michi- 1978, Sexuality and the compon- gan, Museum of Zoo1 ogy , Mi scel- ents of environmental uncertain- laneous Publication 156, 124 p. ty: clues from geographic par- thenogenesis in terrestrial ani- Hubricht, Leslie, 1960, The cave mals: American Naturalist, v. snai 1 , Car ychi um sgyql um Call : 112, p. 659-673. Transactions Kentucky Academy of Science, v. 21, p. 35-38. Hassell, M. D., and Harvey, M. J. , 1965, Differential homing in Hubricht, Leslie, 1962, New spe- Plyotis soda1 is: Amererican Mid- cies of Helicodiscus from the lands Naturalist, v. 74, p. 501- eastern United States: Nauti 1 us, 503. v. 75, p. 102-107.

Hobbs, H. H., Jr., and Barr, T. C. Hubricht, Leslie, 1963, New 5pe- Jr., 1972, Origins and affini- cies of Hydrobiidae: Nautilus, ties of the troglobitic cray- v. 76, p. 138-140. f i shes of North America (Deca- poda: Astacidae). 11: Genus Hubricht, Leslie, 1954, Land conectes: Smithsonian Contribu- snai 1s from the caves of Ken- tions to Zoology, v. 105, 84 p. tucky, Tennessee, and Alabama: Bull et i n of Nat i onal Spel eol og- Hobbs, H. H., Jr., Hobbs, H. H., ical Society, v. 26, p. 33-7,56. 111, and Daniel, M. A., 1977, A revi ew of the troglobitic deca- Hubricht, Leslie, 1965, Four new pod crustaceans of the Americas: land snails from the southeast- Smithsonian Contributions to ern United States: Nautilus, v. Zoology, v. 244, 177 p. 79, p. 4-7.

Holsinger, J. R., 1965, Redescrip- Hubricht, Leslie, 1968a, The land ti ons of two poor1 y known spe- snai 1 s of Mammoth Cave Nati onal cies of cavernicolous rhagidiid Park, Kentucky: Nautilus, v. 82, mites (Acarina: Trombidif ormes) p. 24-28. from Virginia and Kentucky: Acarologia, v. 7, p. 654-662. Hubricht, Leslie, 1968b, The land snails of Kentucky: Sterkiana. Holsinger, J. R. , 1978, System- 32, p. 1-6. atics of the subterranean amphi- pod genus S5yqobromus (Crango- Jeannel, Rene, 1949, Les coleop- nyctidae), part 11: Species of teres cavernicoles de la region the eastern United States: des Appalaches. Etudes system- Smithsonian Contributions to atiques: Notes Biospeol., fasr, Zoology, v. 266, 144 p. 4, Publ. Mus. of Nat. Hist. Nat., Paris, no. 12, p. 37-104. Holsinger, J. R., in press, Zooge- ogeographic pattern in North American subterranean amphipod Kenk, Roman, 1977, Freshwater tri- I crustaceans, jn F. R. Schram, clads (Turbellaria) of North . I ed., Issues 4, Crus- America, IX: The genus Sphalz tacean : Rotterdam, loplana: Smithsonian Contribu- A. A. Balkema. ------tions to Zoology, v. 246. 38 _el ~ Hubbell, T. H., and Norton, I?. M. , EreknLerl-Gari-H,I-L97311-C:a_~e_-b_e_e_~ 1978, The systematics and biol------tles of the genus ---- Pseudanoph- ogy of the cave-crickets of the ------thal mus (Coleoptera, Carabi dae) North American tribe Hadenoecini from the Kentucky Blue Grass and ( Saltatoria: Ensi- vicinity: Fieldiana (Zoology), fera: Rhaphidophoridae: Dolicho- v. 62, p. 35-83. I

CAVE LIFE OF KENTUCKY 167

University , Museum of Compara- enden Thiere: Mullers Arch. f. tive Zoology, Bulletin, v. 144, Anat. u. Physiol., v. 4, p. 384- p. 151-352. 394.

Swofford, D. L., Branson, B. &. , Tellkampf, T. G., 1845, Memoirs on and Sievert, G. A. , 1980, Genet- the blind fishes and some other ic differentiation of cavefish animals living in Mammoth Cave populations: Isozyme Bulletin, in Kentucky: New York Journal of v. 14, p. 109-110. Medicine, v. 1845, p. 84-93.

Valentine, J. M., 1952, New genera Tellkampf , T. G. , 1844a, Beschrei- of anophthal mid beet1es from bung einiger neuer in der Mam- Cumberland Caves (Carabidae, muth-Hohle in Kentucky aufge- Trechinae): Geological Survey of f undener Gattungen von Gl ieder- Alabama Museum Paper 34, 41 p. thieren: Arch. f. Naturg., v. 10, p. 318-322. Woods, L. P., and Inger, H. F., 1957, The cave, spring, and Tellkampf , T. G. , 1844b, Ueber den swamp fishes of the family Am- blinden Fisch der Mammuth-Hohl e blyopsidae of central eastern in Kentucky, mit Bemerkungen United States: American Midlands uber andere in dieser Hohle leb- Naturalist, v. 58, p. 232-256. 168 Chapter 10 VERTEBRATE REMAINS IN KENTUCKY CAVES Ronald C. Wilson Director, University of Northern Iowa Museum University of Northern Iowa Cedar Falls, Iowa 50614

One of the earliest vertebrate time. The potential of paleontolo- fossils to be recognized in North gical work in Kentucky caves has America (preceded only by those at hardly been scratched. Pig Bone Lick, Boone County, Ken- tucky) came from a Kentucky cave. WHY STUDY BONES? A skull of a f lat-headed peccary -----(Platyqonus ------compressus) ------was col- Cave bone deposits preserve a lected by Dr. Samuel Frown, about record of past animal communities 1804, from Great Saltpeter Cave, spanning a range of at least sev- Hockcastl e County. The specimen is eral tens of thousands of years. still in the collection of the These deposits represent an oppor- Philadelphia Academy of Science. tunity to study the paleoecology Despite this early interest in the of many extinct species. Ey vertebrate paleontology of Ken- inference from known geographic tucky caves, the area has received ranges and ecol ogi cal to1erances little study since. During the of 1iving species preserved in nineteenth and early twentieth cave deposits, it is possible to centuries, only the spectacular reconstruct the evolution of finds (e. g. , complete sku1ls, changing climates and environments mammoth or mastodon bones, etc.) in the Ohio Valley. Occasional 1y, were likely to be recorded. These we1 1 preserved archeological /pal e- ear ly discover ies , however, were ontological sites permit recon- likely to be lost or forgotten be- struction of the activities of fore any scientific study unless prehistoric man and other preda- the material went to the large tors of the distant past. Perhaps museums in the East. Host modern most importantly, the kinds of scientific studies in the verte- information preserved in cave bone brate paleontology of Kentucky deposits are frequently not pre- were accomplished under the direc- served in any other paleontolo- ton of the late John E. Guilday gical context. If bone deposits (Carnegie Museum of Natural are destroyed or disturbed by History) during the 1960's and natural or human processes, part early 1970's. As his student, of the prehistory of an entire I express my gratitude to John for region is lost forever. his training and encouragement. Anal ysi5 of bones from several HOW DO BONES GET INTO CAVES? recently excavated sites is still in progress. Some of the new finds Cavers are frequently surprised are reported here for the first to learn that remains of large VERTEBRATE REMAINS IN KENTUCKY CAVES animals are found deep inside utilization and the degree of caves, many kilometers and hours mineral izaton of the bones from the nearest entrance. present. Travertine or gypsum may Recovery of the bones of giant encrust fossil bones. Entrances short-faced bear (Arctodus ~imus)~ may collapse or fill in, altering flat-headed peccary (Platygonus or ending their accessibility to com~ressus), and American various kinds of animal activity. mastodon (Mammut americanum) or These processes may produce mammoth (Mammuthu~Lfrom Proctor extremel y complex stratigraphy and Cave in 1979, for example, require careful excavation and required 24-hour cave trips. The record keeping to decipher. mastodon or mammoth bones were Gnawing and digging by animals found in the talus of a terminal such as woodrats, woodchucks, and breakdown while the remains of the carnivores may further confuse the other species were scattered along record. In Kentucky, those an active cave stream. They had deposits not buried soon after evidently been washed to that side deposition are of ten so completely by floods after being eroded from gnawed that the accumulation is the site of original deposition reduced to very small bones such farther upstream. as mouse and parts, and the Most bones are deposited near harder crowns of large present or past entrances. teeth. Complete and articulated . Exceptions to this general skeletons are extremel y rare 'in statement may include bones that Kentucky caves. wash in, or skeletons of animals that wandered far from the WHAT HAS BEEN FOUND? entrance before dying. Vertical Most caves contain bones. Some shafts act as pit-f all traps and contain extensive deposits, they are responsible for representing thousands of innumerable cave bone deposits. individual animals and dozens of Other common means of bone species, while others may contain accumulation include the on1y an occasional scrap of bone. activities of woodrats, The significance of any particular carnivores, and raptors (predatory bone deposit depends on a variety such as hawks and owls). Man of factors including the age, has been an agent in the number, diversity of species accumulation of bones in Kentucky represented, quality of caves for at least the last 10,000 preservation, presence and number years. Detailed analysis is of ten of extinct or extralimital (no required to determine the mode of longer living in the local region) accumul ati on, and many bone species, and the taphonomy deposits are the composite of (history of accumulation and several processes. subsequent alteration) of the During the thousands of years deposi t . that an entrance may be open for Species represented in cave animals to enter, geolgic bone deposits are either extinct, processes may have significant still living in the area of the effects on the resulting bone cave, or eetralimital. Those deposi ts. Roc kf a1 1s and species still living in the sedimentation may seal layers of vicinity of the cave seldom bones from contamination by more provide dramatic insights into recent accumulations. Erosion may changing environments, but they expose older layers as drainage can serve as reminders of the patterns shift, and redeposition relative adaptability of some of the f ossi 1s may occur. species. They may also provide Fluctuations in the water table information on when additions to affect not on1y sedimentation the local fauna first arrived in a patterns, but a1so ani ma1 particular region. 170 WHAT HAS BEEN FOUND?

Indications of how climates and 12,950 + 550 years before environments change may be present) and Toolshed Cave, inferred from interpretation of Eullitt County, numerous f ossi 1s of extralimital species. A individuals were recovered that few examples will serve to may represent herds. At least 31 indicate the significant changes individuals were recovered from in the Kentucky mammalian fauna. the Welsh Cave deposits and at Northern species like porcupine 1east two dozen from Tool (Erethizon doprgtum), red squirrel Cave. Flat-headed peccaries are (Tamiasciurus hudsonicus), heather common in Kentucky caves because or spruce vole (Phenacomys), they traveled in herds and, like yellow-cheeked vole (Microtus 1i vi ng peccary species, they of ten xanthoqnathus), snowshoe hare used caves as shelter. Typical 1y, (Leeus americanus), wolverine the fossil herds are composed of ------(Gulo qulo), pine marten (Hartes young animals on1 y a few weeks old ameri canal and 1east at the time of death as well as ------(Mustela nivalis) have been found young adults and aged individuals in Kentucky caves. A 1920s report identifiable by their worn teeth by a University of Kentucky and advanced arthriti s. graduate student of a PLatyg~n~s-ve&us,a larger ...... (Thalarctos maritimus) is almost species that preceded P, certainly based on a Gpgpressus evol uti onar i1 y , has misidentification. Unfortunately, been found in Lisanby Cave, the specimen has been lost. Caldwell County. Discovered by Western species recovered from cavers who were mapping the cave, Kentucky caves include 13 lined this may be the oldest Kentucky ground squirrels (Speymophilus bone deposit discovered to date. tridecemlineatus), plains pocket Although precise dating is not (Geomys bursarius), Mexican possible, these bones are probably f ree-tai led bat (Tadarida more than 250,000 years old. 6 brasilien~is)~ (Taxidea third species, the long nosed taxus)%and grizzly bear (Ursus peccary, Eylphyus-nazutuz, has acrtos). Kentucky caves also been recovered in Savage Cave, contain evidence of the former Logan County, and Icebox Cave, presence of many species that have Be11 County. Mylohyug traveled disappeared (or nearly so) from alone or in small f ami 1y groups the local fauna during the and it preferred more wooded areas historic period. These include elk than Platygonus. or wapiti (Cervus ela~hus), gray Horses evolved in North wolf (Canis lupus), red wolf JC, America, but became extinct at the ni~er), mountin lion (Feli~ end of the Pleistocene (Ice Age) concolor), black bear (Ursus about 10,000 years ago. A related americanus) % clutra European species was introduced by canadensis) % prairie chicken Spanish explorers in the sixteenth --(Tympanuchus cupido)%and whooping century. Bones or teeth of an crane (Grus americana). extinct horse species have been Extinct species reveal the found in several caves in the Blue 1imi ts of environmental Grass Region of the State. TWO flexibility and are perhaps the species of tapirs, small relatives most intriguing of Kentucky of the horse, with long flexible vertebrate fossils. The extinct snouts, have also been preserved species that have been found in in Kentucky caves. The smaller Kentucky caves are listed in Table Tapirus veroensis was found in the 1. The most common species is the late 1970's in Proctor and Bowman flat-headed peccary, having been Saltpeter caves. The 1arger found in at least 8 Kentucky --Tapirus ------haysii --- was recovred early caves. In the case of Welsh Cave, in this century from a sinkhole in Woodford County (Carbon-14 date: Scott County . VERTEBRATE REMAINS IN KENTUCKY CAVES

Table 1. Extinct recovered from Kentucky caves.

SCIENTIFIC NAME COMMON NAME SITE NAME

Order Edentata

beaut if ul *A-maze-i n Cave, armadi 1lo Bulli tt Co.

Mesalnsrx-inffersnsii Jefferson '5 Glass Cave, Franklin Co. Gillenwater Cave, Barren Co.

Order

Canis-dL~us dire wol+ Welsh Cave, Woodf ord Co.

giant short-f aced Glass Cave, bear Franklin Co. Proctor Cave, Edmonson Co.

jaguar A-maz e-i n Cave, Eullitt Co.

Order Hodent ia

Cast oroides ohioensi s giant beaver Cutoff Caves, Trigg Co.

Order Perissodactyla

Eguus-cnmelicafus comp 1ex -toothed fissure at Mundy '5 horse Landing , Mercer Co.

!%IUUSSP- horse Glass Cave, Franklin Co. Welsh Cave, Woodf ord Co.

Hay's tapir sinkhole, Scott Co.

Vero Tapir Bowman Saltpeter Cave, Rockcastle Co. Proctor Cave, Edmonson Co.

Order Pirtiodactyla

Mulnhyus-nasutus lonq-nosed Icebox Cave, peccary Bell Co. Savage Cave, Logan co. WHAT HAS BEEN FOUND?

Table 1. (Continued)

rlaf~snnus-vetus Leidy's Lisanby Cave, peccary Caldwell Co. Plafysnsus-cnrnere~sus f 1at-headed Granny Puckett Cave peccary Hart Co . Great Saltpeter Cave, Rockcastle Co. Lone Star Peccary Cave, Hart Co. Proctor Cave Edmonson Co. Savage Cave, Logan Co. *Tool shed Cave, Bullitt Co. Welsh Cave, Woodf ord Co. unknown, Wayne Co. Wells Cave, Boyle Co.

Order Proboscidea

------Mammut americanum Amer ican *Hall's Cave, mastodon Fullitt Co. Tool shed Cave, Bullitt Co. Turner Cave, Barren Co. Walnut Hi 11 Farm Cave, Fayette Co.

mammoth Welsh Cave, Woodf ord Co.

*Investigation of A-maze-i n, Hal 1 '5, and Tool shed caves during 1984 was supported by a grant from the Kentucky Heritage Council to Philip J. DiHlasi and Ronald C. Wilson.

Other unusual species that are Cave, Franklin County and represented by fossils in Kentucky Gillenwater Cave, Barren County), caves include Castoroides a giant ground sloth larger than a ohioens&s (Cutoff Caves, Trigg black bear. County), a giant beaver that Several extinct large approached the size of a black carnivores are also among the bear ; Qglypus be11 us ( A-maz e-i n species recovered from Kentucky Cave, Eullitt County), a giant caves. The largest is the giant armadi 1lo about twice the size of short-faced bear, fiyctodus simus. the 1iving North American species; Larger than a grizzly bear, this and Meqalpnyx jef f ersoni i (Glass giant carnivore has been recovered VERTEBRATE REMAINS IN KENTUCKY CAVES 173

Figure 1. Locations of Kentucky caves where remains of extinct species have been collected.

from Glass Cave, Franklin County disintegrating fyom lack of care and Proctor Cave, Edmonson County. in someone ' s or c 1oset . Other extinct carnivores include the dire wolf, Canis dirus (Welsh Cave, Woodford County) and the .THE CAVER'S ROLE jaguar, Felis onca auwsta The value of vertebrate fossils (a-maze-in Cave, Bullitt County). lies in the opportunities they Discovery of the jaguar in 1984 provide for interpretations of was bittersweet. The first extinct paleobiology and past climates. As large cat to be recorded in in the case of archeology, Kentucky, the bone preservation stratigraphic relationships are was better than in any previous important clues to the accurate find of the species in the eastern interpretation of any fossil United States. The large mature assembl ages. Excavation by cat was crippled by an injured toe untrained individuals should never when it sought the dry shelter of be attempted unless they are A-maze-in Cave. It died curled up working under the guidance of a in a small dry alcove and lay professional . there undisturbed for more than Cavers are responsible for many 10,000 years. When found, however, significant bone discoveries. a1 1 that was left were foot bones, Fossi 1 bone deposits in We1 sh, ribs, sternum, one vertebra, and a Tool shed, Icebox, Lone Star knee cap. These were found Peccary, Cutoff, Lisanby, Granny scattered among the back dirt Puckett, Proctor, and Bowman piles of a looter's pit. Had it Saltpeter caves are all first been recovered intact, it would reported by cavers with no have provided the first exhibit paleontological experience. qua1 it y speci men for museum Discovery of such deposits display as well as biological data requires good observation skills. on the species. Instead, the Most deposits first reveal skull, jaws, vertebrae, and all themselves by a few small major limb bones are lost to fragments of bones or teeth on the science and are probably surface of the cave floor. Bones 174 REFERENCES CITED are more likely to be old if they appropriate comparative material. are stained brown or black by Once identified, bone collections minerals, brittle from loss of belong in repository institutions organic matter due to leaching, where they will be available to preserved in orange, yellow or future researchers and where they light brown seiments or beneath can be maintained with all f lowstone. The major bone deposit accompanying information. Since no in Toolshed Cave, for example, was Kentucky institution at present beneath four 1ayers of f 1owstone, employs a vertebrate but was exposed by erosion of a paleontologist and no collection stream channel. Eones are probably of vertebrate fossils in Kentucky very recent if they are in an institutions are being actively active den or nest, if they are in curated, Kentucky cavers are dark brown sediments, or if they forced to look elsewhere to insure still contain grease. the proper long term management of If bones are to be collected their discoveries. The institution for evaluation by a specialist, with the largest Kentucky the initial sample should be small collections, and a good track and select. It should include record for professional management those elements that have the of collections, is Carnegie Museum highest probability of being of Natural History, Pittsburgh, identifiable: jaws, teeth, major Pennsylvani a. However, other 1imb bones, pelvis or large ankle midwest state museums ( I11 i nois, bones. Bones should not be Indiana, etc.) also have trained collected unless they can be personnel. Further gui dance on packed well enough to survive the dealing with bone discoveries can trip out of the cave. If bones are be obtained by contacting the very distinctive or too fragile to author of this chapter or the NSS collect, they may be identified Vertebrate Paleontology Section. from a photograph or detailed The f ossi 1 bones in Kentucky sketch. The Srctodus manible from caves have just begun to reveal Proctor Cave was first identified their secrets. Based on recent from a detailed sketch in the discoveries in neighboring states, notebook of a Cave Research exciting finds can be expected at Foundaton survey party. any time. Rmong the discoveries to Whether the initial collection be expected are sabertooth cats, of data is a sketch, photograph, camel, cheetah, better preserved or representative col lection of material of other extinct species bones, it must be labeled discussed in this report, and immedi ate1y. The 1abel should additions to the list of include the name and location of extralimital species from the the cave (and of the site within State. These discoveries await the cave if the cave is a large informed cavers who are aware of one), the name of the the possibiliies and who have collector(s), and the date. Once sharp eyes. out of the cave, the bones should be repacked for shipment to a competent faunal analyst or vertebrate paleontologist. REFERENCES bones-shnuld-nsf-be--washedL If breaks occur, be sure to save all Cooper, C. L., 1931, The Pleisto- the pieces and wrap them so they cene fauna of Kentucky: Kentucky do not rub against each other Geological Survey, Geologic Re- during shipment. port, v. 36, p. 435-460. Eones can be reliably identified only by professionals Guilday, J. E., Hamilton. H. W., with extensive training, and McCrady. A. D. , 1971, The experience and access to Welsh Cave peccaries (Plsty= VERTEBRATE REMAINS IN KENTUCKY CAVES 175

qpngsL and associ ated fauna , Miller, A. M., 1923, Recent cave Kentucky Pleistocene: Anna1 s explorations in Kentucky for an- Carnegie Museum, v. 43, no. 9, imal and human remains: Kentucky p. 249-320. Geological Survey, ser. 6, v. 10, p. 107-113. Gui lday, J. E. , and Parmalee, P. W., 1979, Pleistocene and Webb, W. S., and Funkhouser, W. D. recent vertebrate remains from 1934, The occurrences of the Savage Cave (15Loll1, Kentucky: f ossi 1 remains of Pleistocene Western Kentucky Speleological vertebrates in the caves of Survey Annual Report, 1979. Barren County, Kentucky: Murray, Kentucky, p. 5-10. Lexington, Kentucky, University of Kentucky Reports in Archeol- Harris, Arthur H., 1976, Paleon- ogy and Anthropology, v. 3, no. to1ogy: National Cave Management 2, p. 39-65. Symposium Proceedings, 1975, p. Wilson, H. C., 1981a, Extinct ver- brates from Mammoth Cave, jn Hay, Oliver P., 1923, The Pleisto- Proceedings of the Eighth Inter- cene of North America and its national Congress of Spel eol ogy: vertebrated animals: Washington, p. 339. Carnegie Institute, pub. 322. Wilson, R. C., 1981b, Preliminary report on vertebrate remains Jegla, T. C. and Hall, J. S., from Cutoff Caves, Trigg County, 1962, A Pleistocene deposit of Kentucky, Western Kentucky the free-tailed bat in Mammoth Speleol ogi cal Survey Annual Re- Cave, Kentucky: Journal of port, 1980: Murray, Kentucky, p. Mammallogy, v. 43, no. 4, p. 35-37. 477-48 1. Wilson, R. C., 1982, The recogni- tion, evaluation, and management Jillson, W. R., 1968, The extinct of cave bone deposits, &n vertebrata of the Pleistocene National Cave Management vertebrata of the Pleistocene in Symposium Proceedings: 1978, Kentucky: Frankfort, Kentucky, 980, p. 121-122. Roberts Printing Co., 122 p. Wilson, R. C., Guilday, J. E., and Pranstetter, J. A., 1975, Ex- Kurten, E. and Anderson, E., 1980, tinct peccary from a central Pleistocene mammals of North &m- Kentucky cave: National Speleo- erica: New York, Columbia Univ- logical Society Bulletin, v. 37, ersity Press, 422 p. p. 83-87. Chapter 11 ARCHEOLOGY Patty Jo Watson Department of Anthropology Washington University St. Louis, Missouri 63130 and

The best cavers in the world-- tural entrances, the aboriginal unti1 about A. D. 1950--were some human inhabitants entered caves of the people who lived in the in search of exploitable resour- area of Mammoth Cave between 2,000 ces, to contact supernatural for- F. C. and O B. C. /A. I?. Baref eoted, ces, or just to satisfy their cur- or shod with thin woven-f iber iosity. This chapter is a brief moccasins, and very lightly clad summary of what is known about the otherwise, prehistoric Indians archeology of Mammoth Cave and of used dry weedstalk or cane torches several other Kentucky caves and to light their way through the karst features, making comparative main trunk to lower-lying canyon reference to archeological remains passages and crawlways several in Wyandotte Cave in Indiana and kilometers from the cave entrance in some Tennessee caves. (Fig. 1). They explored and mined cave minerals extensive1y in several parts of the Mammoth Cave System, especially Salts Cave and Mammoth Cave. At least two of them died in the cave: one was a boy no more than 9 years old, the other a man of about 45 (Pond, 1937: Neumann, 1938: Meloy and Watson, 1969; Robbins, 1971, 1974: Meloy, 1984). What we know about their lives and their caving activites is summarized in a number of pub- lications (Watson and others, 1969: Watson, 1974, 1984; Stein and others, 1981).

For several years it was thought that the archeological remains in the Mammoth Cave System were unique, but we now know that to be false. It appears that throu~hout- the entire mid- continental karst region of the United States, wherever there Figure 1. Salts Cave. Imitation were re1ativel y accessible na- aborigine. CRF photo. ARCHEOLOGY

THE MAMMOTH CAVE SYSTEM

The first comprehensive attempt to assess Mammoth Cave area pre- history was made by the archeol- ogist Doughas Schwartz in the late 1950's when he was on the faculty at the University of Ken- tucky (Schwartz 1960, 1965). Schwartz's work was preceded by that of N. C. Wilson (1917) and Alonzo Pond t 1937) , a.nd has been succeeded by that of the Cave Research Foundation Archeological Project beginning in 1962 (Eenington and others, 1962; Figure 2. Narshall Avenue, Lee Stein and others, 1981: Watson Cave (Joppa Ridge, 1;;armoth Cave and others, 1969; Watson and National Park). Torch canes. Yarnell, 1966: Watson, 1974; CRF photo, Pete Lindsley. see also CRF Annual Reports from 1977 to the current year). fecal deposits ( paleof eces) that On the basis of that body of contain invaluable dietary research, one ran suggest the informatisn (Figs. 3 and 4). A f 01 lowing interpretation of Mammoth Cave prehi story. The inhabitants of the region now comprising Mammoth Cave National Park began exploring at least the upper trunk passages in Mammoth Cave about 4,000 years ago Cca. 2,000 B.C.). Between that period and 0 B.C. /A.D. they freely expl ored several ki1 ometers of cave passages of a1 1 sorts, including some very low crawlways (for example, Wilson's Way off Ganter Avenue in upper Mammoth Cave) . Nature of the Archeological Remains In spite of 175 years of recent traffic in the dry, upper levels of this great cave, abundant--if scattered and high1y f ragmented--aborigi nal materials remain as testimony to the wide-ranging and persistent aboriginal presence. For the most part, these are fragments of torch and campfire fuel: cane, dried weed stalks, twigs, and branches (Fiq.- 2). But there are also pieces of cordage, portions of vegetable-fiber moccasins, broken Figure 3. Climbing pole, possi- bowls made of gourds or of bly prehistoric, in the Upper thick-walled gourd-like squashes, Trunk Passage, Mammoth Cave. and dozens of prehistoric human CRF photo, William McCuddy . 178 NATURE OF THE ARCHE OLOGICAL REMAINS

Figure 4. Ganter Avenue, Mammoth Cave. Basket, probably prehistoric. CRF photo, Roger Brucker . summary of radiocar-bon dates on many of the materials is shown in Table 1. Other indications of prehistoric exploration are charcoal smudges on walls, ceilings, and breakdown blocks, as well as batter marks where thp gypsum crust was hammered aff passage walls (Fig. 5), perhaps to Figure 5. Salts Cave. Gypsum be used in making white paint. crust battered off wall (top) , There are also many places* mirabilite and gypsum crystals especi a1 1y in upper Mammoth and on ceiling of a little ledge upper Salts, where sediments have (center), and torch smudges on been dug away in search of other wall below ledge (bottomJ. CRF crystal line forms of gypsum photo. !satinspar and selenite). A1 though the evidence is iifficult to interpret, most work stored or cached in the cave on a parties were probably small. A temporary basis) (Fig. 6). The very eificient size would be 6 author believes these prehistoric people, with 2 serving as more or cavers were as comfortable in the less f ul1 -ti me torch bearers when cave a5 are contemporary cavers, the group was moving and and--given the limitations of fire-tenders when they were their equipment --were just as working in one place. Experiments competent at subterranean with cane torches (Ehman, 1966; navigating, climbing, and hiking. Watsan and others, 1969, p. 60-52) Of course, we cannot be sure we have shown that they are a very know everything about why these zatisf actory form of cave 1ight , prehistoric people went and that it is quite possible to underground because some of their carry enough dry cane to iast many activities may not have left any hours (10 or 12, or even more if material remains. Eut it is clear several people carry spare fuel. that they frequently sought and Torch fuel may also have been obtained the various forms of ARCHEOLOGY

MAMMOTH CAVE NATIONGL PARK Prehistory: Summary o+ Available Radiocarbon Dates (Libby half-life, uncaiibrated)

PROVENIENCE MATERIGL DATE ._...... ______saLts-Gsve-Yezfib~!Le Test 11. F. 2A Charcoal 1540 B.C. + 110 (GaK 27673 Test E, Level 5 Charcoal 1410 B.C. 2 220 (GaE 2764) Test G, Level 5 Charcoal 1460 B. c. + 220 (Gat.-:: 2766) Test E, Level 7h Charcoal 710 B.C. 2 100 (GaK 2622) Test E, Level 7b Charcoal 990 B.C. t 120 (Gak 2765) Test JIV, Level 4 Charcoal 52(:, B.C. + iO Test J I Ci , Level b Charcoal 390 B. C. + 50 Test JIV, Level 8 Charcoal 480 B.C. + 50 Test J I 'b' , Level 11 Charcoal 560 B.c. + 60 Test KII 7 Level 4 Charcoal 570 B.C. + 70 f est KII, Level 6 Charcoal 250 B.C. + 60 Test KII, Level 11 Charcoal 431:) B.E. + br:, Test KII. Levei 14 Charcoal 460 S.C. + &[I) salts-Cave-InterLor Upper Salts, F54 Paleo-Fecal specimen 290 5. C. + 200 with ssuash seeds (M 1573) Upper Salts, P38 Paleof ecal specimen 320 B.C. 2 140 (M 1777) Upper Salts, P63-64 Paleof ecal specimen 620 B.C. + 140 with gourd seeds (M 1574) Upper Salts, P54 Soot 1125 B.C. 140 (I 256)

Upper Salts, Test A 0-10 cm Cane B.C. 2 140 1584) 30-40 cm Cane P.C. + 130

1585) a 70-80 cm Cane B.C. _+ 150 1586) 140 cm Wood B.C. 2 140 1587)

Middle Salts, Blue Arrow Passage. Paleof ecal specimen 400 B.C. + 140 460 with squash polien (M 1577) A42 Paleo+ecal specimen 710 B.C. 2 140 with sun+lower (M 1770) achenes Lower Salts, Indian Avenue, 176 Wood 770 B.C. 2 140 (M 1588) I67 Wood and bark 1190 B.C. 2 150 (M 1589) 180 NATURE OF THE ARCHEOLOGICAL REMAINS

Table 1. (Continued)

PROVENIENCE MQTERIGL DATE

Salts Cave Internal tissue A.D. 30 2 160 (M 2258) Internal tissue 10 E.C. + 160 (M 2258) Mammeth-Cave-Lnterl~c Upper Mammoth Sl ipper 28Cl B.C. f 40 (X 8) Cane 420 E.C. + 60 ix 9) Lower Mammoth Ganter Avenue. B10 Wood 1050 B.C. 2 70 (UCLA 1730B) 2170 R.C. 2 70 iUCLA 1730G) Mammoth Cave Mummy matt in^ 445 E.C. t 75 (51 30t37Q) Internal tis~ne 15 B.C. _+ 65 (SI 3007C) LEE-cave-Lntcrioc Marshall Avenue. KBS Cane

gypsum noted above. Medicinal it is very likely that these were sulfate sal ts (most common1 y also of interest. Both epsomite epsomite and mirabilite) also and mirabilite are excellent occur naturally in the caves, and cathartics, and mirabilite (being a form of sodium sulfate) is also salty, so both were probatclv mlned as well as the gypsum.

Aboriginal Caving Techniques Techniques were simple but effective insofar as one can tell from the evidence remaininq. No special clothing was worn, and In fact there are several we1 1-preserved footprints in dust or mud (Watson and others 1969, p. 63, Plate 14) that indicate the Indians at least sometimes went barefoot in the cave (Fig. 7). It is quite possible to negotiate crawl ways and breakdown cli mbs Figure 6. Salts Cave. Bundle while hnlding a torch in one !-land. of twigs and sticks, apparently Chimneying and canyon straddling cached in the breakdown and lost wauld have been a greater or forgotton (note torch smudges challenge but by no means on rock above the bundle) . CRF impossible, especi a1 1y with photo, Robert Hall. cooperation among torch-bearers ARCHEOL OGY 181

Figure 7. Lower Salts Cave. Canyon passage explored by the Indians. CRF photo, Mark Elliott.

and non-torch-bearers (Fig. €I). Two or three cane torches will light the largest rooms and passages much better than carbide 1amps or even battery-powered headlamps (because the light is more evenly diffused). Hence, a party of 8 or 10 could move about with ease. as long as they kept fairly close together, using the Figure 8. Mummy Valley, Salts light from only 2 or 3 torches. Cave, with imitation aboriginal Several people would then have explorers. CRF photo, James Dyer. their hands free to carry food, water, collecting bags or other containers, and spare fuel. Mammoth Cave System is that of how The torches themselves were the Indians made a living. Traces several pieces (3 to 5 seems of their food are very well optimum) of cane 2 to 3 feet long, preserved in the dried excrement or several weed stalks of simi lar that is still present in many convenient length. These torches places. We know they were growing were sometimes loose1y bound some plant foods such as together with strands of inner sunflower, and were harvesting the bark fiber. The techniques for nuts and fruits from several making fire are not known, but forest species such as hickary, could have included various oak, blackberry, and strawberry. systems involving applied They were also growing gourds and friction. Twirling a hardwood squashes? both of which were stick against another, flatter probably used as containers (these piere of wood with punk of timber people did not make pottery). placed to catch fire from the although they did sometimes eat friction is one such possibility. the seeds of both plants (Fig. 9). Squash and gourd are a+ special Prehistoric Subsistence interest because they are tropical One of the most productive plants that were first avenues of inquiry stemming from dnmesticated in Mexico or the archeological remains of the somewhere even farther south. 182 ARCHEOLOGICAL REMAINS IN OTHER CAVE AND KARST FEATURES

ARCHEOLOGICAL REMAINS IN OTHER CAVE AND KARST FEATURES NEAR MAMMOTH CAVE NATIONAL PARK Both inside the Park and outside it, prehistoric materials were once present in sandstone and limestone rock shelters. Nearly all of these materials have been badly disturbed by vandals and relic collectors, many of the sites being so completely ransacked that virtual1 y no q.., cante::tual xn+ormation 1s left. These shelters once provided seasonal homes for small groups oi prehistoric people, families or extended f ami 1ies, and temporary campsites for parties of hunters and gatherers. The fragmentary remains in those shelters that Figure 9. Warty squash bowl and have been examined by torch canes in Indian Avenue of archeologrsts indicate Lower Salts Cave. CRF photo, intermittent occupation from Robert Keller. several thousand years ago to a few hundred years ago in various of them. The ancient Kentuckians also However, by 5,000 B. C. they had camped near natural features liKe been traded as far north as the Mi11 Hole, a karst window or lower Illinois River Valley resurgence point south of Mammoth (Conrad and others, 1984). =a it Cave National Park, where they is not surprising to find them in quarried chert from outcrops in the Mammoth Cave part of the Ohio the limestone to be made into River drainage (the Green River, projectile points and other tools. which runs past Mammoth Cave, is Crump's Cave, which opens of+ a one of the tributaries of the sink near Smith's Grove, is Ohio). Nevertheless, the abundance another nearby, karst-related and excel lent preservation of prehistoric campsite (now bad1y botanical remains in the cavern disturbed). In Short Cave, a passages makes them a unique and number of buria1 s--probabl y late extreme1y valuable storehouse of prehistoric--were dug up in the information on plant use and early early 1800s by saltpeter miners. cultivaton in the Eastern On Prewitts b.lnob near the highway Woodl ands . between Cave City and 61asgow, Animal bones--cracked, cut, prehistoric people quarried chert broken, and sometimes charred--are and also disposed of some oi their present in midden deposits once dead in pits in vertical shafts abundant in the entry areas of occurring at the edge of the both Salts Cave and Mammoth Cave. sandstone capping the knob. fi few These bones, together with the footprints and torch remains have much less readily identi-fiable been found in Fisher Ridge Cave animal remains in the fecal near the town of Horse Cave: on deposits. indicate that deer, the basis of two radiocarbon turkey. , opossum. determinations on the torch squirrel, , and other remains, these footprints date to animals, including birds and fish, about 3,000 years ago. The karst were hunted and eaten. window entrance of a cave near ARCHEOLOGY 183

Bowling Green (48 km south of Cave CAVE ARCHEOLOGY ELSEWHERE IN City) was a long-term campsite and chert like Mi11 Hole; THE MIDWEST AND MIDSOUTH unfortunately it has been severely Archeolo~icalremains deep in vandalized in the past 2 years. caves. and thus somewhat Another archeological site in a comparable to the Mammoth Cave cave opening off a sinkhole situatinn, are known from (Savage Cave, in south-central Wyandotte Cave in southern Indiana Kentucky) has been recently and from several Tennessee caves. purchased by the Archeological Middle Woodland Indians explored Conservancy, and is being managed Wyandotte Cave and mined chert and by Murray State University so that aragonite there, which was rather what remains of the cultural wide1 y traded throughout the deposit will be protected. Midwest. In Zarathustra Finally, recent work in the (Saltpeter ) Cave in northern vicinity of Louisville by Tennessee, chert was also mined in paleontologist/zoologist Ron some quantity, and Jaguar Cave Wi1 son (Universi t y Museum. (northern Tennessee) ( 2ee Hobbi ns Uni versi ty of Northern Iowa, Cedar and others, 1981) (Fig. 111, and Falls) and archeologist Phil (central Tennessee) DiBl asi (Archeological Survey. were both rather thoroughly University of Louisvi 1le) has explored prehistorical1 y. The most resulted in the locating of unusual archeological remains from several caves containinq historic any cave are probably those in Mud and prehistoric cultural remains Glyph Cave in eastern Tennessee. (Fig. 10). where drawings in the mud that

Figure 10. Salts Cave vestibule excavations. CRF photo. 184 SUMMARY

Figure 11. Jaguar Cave, near the Tennessee/Kentucky border. ClW photo, William McCuddy.

covers walls of a 100 m long SUMMARY passage have been dated to the Frehistoric people throughout last prehistoric period a few the Midwest and Midsouth made use hundred years ago iFaul kner and of rack shelters and cave others, 1984). entrances as habitations for Mud Glyph is the best known thousands of years. It is nQw representative of what seems to be abundantly clear that as iong ago ritual or ceremonial caves. Very as 2,000 F.C. they also frequently few of these are known at present, entered caves to extract natural but this kind of cave-related resources that were of interest to activity seems to be later in time them. They went deep into the than use of caves and rock interior of the longest cave in shelters for habitation purposes, the world, and at least part of and the exploring and mining of the time some of them were simpi y Mammoth Cave. Little is known reconnoitering, or just plain about the specific ceremonies that caving. During the latter part of might have taken place in Mud the prehistoric period, some caves Glyph Cave, but it is apparent apparently became sacred places that creatures known from historic where ceremonies were held, southeastern Indian mythology are possibly as a means of represented, as are motifs of the communicating with the so-called Southern Cult that subterranean world of the spread through the eastern part of supernatural. North America in the late Archeological remains in caves, prehi storic-protohistoric period. whatever their nature, are ARCHEOLOGY 185 extremely valuable and extremely Kentucky: Illinois State fragile. Should you find anything Museum, Report of Investiga- in a cave that appears to be tions 16, p. 65-69. evidence for historic and prehistoric activity there, do not Nelson, N. C., 1917, Contributions touch it or disturb it. Record as to the archeology of Mammoth much information about it as you Cave and vicinity, Kentucky: An- can and report it to the State thrnpol ogi cal papers of the Archeologist. Most of the recent Rmerican Museum of Natural His- evidence of prehistoric deep-cave tory, v. 22, p. 1-73. utilization has been discovered by cavers rather than archeologists. Neumann, Georg, 1938, The human Only if cavers maintain this remains from Mammoth Cave, Ken- esemplary tradition can our mutual tucky: American Antiquity, v. A,- knowledge of the world underground p. 339-353. continue to increase. Pond, Alonzo, 1937, Lost John of Mummy Ledge: Natural History, v. 39, p. 176-184.

REFERENCES Robbins, Louise, 1971, A Woodland "Mummy" from Salts Cave. Ken- Benington, F. M., Melton, Carl, tucky: Qmerican Antiqui ty, v. and Watson, P. J., 1962, Carbon 36, p. 200-206. dating prehistoric soot from Robbins, Louise 1974, Prehistoric Salts Cave, Kentucky: American people of the Wammoth Cave area, Antiquity, v. 28, p. 238-241. --in Watson, Patty Jo, ed., Arche- ology of the Mammoth Cave area: Conrad, Nicholas, Asrh, , New York, Academic Press, p. Asch, Nancy, Elmore, David. 137-162. Gove, Harry, Hubin, Mayer, Brown, James, Wiant, Michael , Farnsworth, Kenneth. and Cook. Robbins, Louise, Wilson, 13. C.; Thomas, 1984, Accelerator radio- and Watson, P. J., 1981, Palean- carbon dating of evidence for tology and archeology of Jaguar prehistoric horticulture in Cave, Tennessee: Eighth Intern3- Illinois: Nature, v. -a08 p. 443- tional Congress of Spel eol ogy, 446. Proceedings, v. 1, p. 377-ZBC1.

Ehman, M. F., 1966, Cane torches Schwarti, D. W., 1960, Prehistoric as cave i 1lumination: National man in Mammoth Cave: Scientific Speleological Society News, v. American, v. 203, p. 130-140. 247 p= 34-.36. Schwartz , D. W. , 1965, Prehistoric , C. H., Deane? B., and man in Mammoth Cave: Eastern Na- Earnest, Howard, 1984, A Missis- tional Park and Monument As5~c- sippian period ritual cave in iation, Interpretive Series No. Tennessee: Amer i can Anti quity. 2, 10 p. Meloy, Harold, 1984, Mummies of Mammoth Cave: Shelbyville, Indi- Stein, J.: Watson, P. J.; and ana, Micron Publishing Co. White, W. 6. , 1981, Geaarcheol- ogy of the Flint Mammoth Cave Me1 oy, Harold, and Watson, P. J. , System and the Green River, 1969, Human remains: "Little Western Kentucky: Geological So- Alice" of Salts Cave and other ciety of America, l9Bl Annual mummies, in Watson and others, Meeting, Cincinnati, Ohio, The Pre-hist~ryof Salts Cave, Guidebooks, v. 111, p. 507-542. 186 REFERENCES CITED

Watson, P. J., ed., 1974, Arch- 1966, Archeological and paleo- eology of the Mammoth Cave area: ethnobotanical investigations in New York, Academic Press, 255 p. the Salts Cave National Park, Kentucky, American Antiquity, v. Watson, P. J., 1984, Ancient In- 31, p. 842-849. dians of Mammoth Cave: Science Year 1984: World Book Science Watson, P. J., and others, 1969, Annual, p . 140- 153. The prehistory of Salts Cave, Kentucky: Illinois State Museum, Reports of Investigations, No. Watson, P. J., and Yarnall, H. A.. 16. Chapter 12 CAVES AND THE SALTPETER INDUSTRY IN KENTUCKY Stanley D. Sides Cave Research Foundation 2014 Beth Drive Cape Girardeau, Missouri 63701

Cave exploration in the pioneer of 1812. Unfortunately, this may days of the State of Kentucky was not be correct. Its use in gun- not for recreation, commercial powder made supply and demand for tourism, or scientific study, but saltpeter vary greatly in times of rather for the mining of minerals war. This war-related inflation of necessary for survival . Common value led to intensive exploration salt, sodium chloride, was almost of the caves of Kentucky, for they the only preservative early set- were an important economic re- tlers had to pickle or cure their source. The early interest in the beef or pork, and preserve their caves, nearly 200 years ago. has game. Natural brines from salt left a rich history for modern- licks were so important that Con- day historians and speleologists. gress encouraged development of There are important para1lels salt springs wherever they were between the development of the found. saltworks of Kentucky, and the Also important as a chemical to saltpeterworks. The technology of pioneer survival was potassi um making salt from brines and making nitrate, or saltpeter. The name saltpeter from calcium nitrate "saltpeter," with its various leached from earth was similar. spellings, meant "salt of earth" Prominent salt merchants also sold or "salt of rock," in contrast to saltpeter. Without these chemical common salt, which came from industries, the settlement of water. Calcium nitrate deposits in Kentucky might have occurred caves and rock shelters were differently. eagerly sought, and an industry was developed to convert the raw mineral to purified saltpeter. SALTPETER AND EARLY Saltpeter is an important chemical KENTUCKY HISTORY for the preservation of meat. Daniel Boone first settled in Medicinally, it is used as a Kentucky in 1769, a period of diuretic. Saltpeter is also a exploration and settlement of the necessary component, with sulfur Blue Grass Region which left the and charcoal , for the manufacture pioneers far from sources of of gunpowder for firearms, fuses, chemicals in Virginia. To the west and blasting. of Kentucky, settlements in what Histories of saltpeter mining are today Missouri and Illinois in Kentucky dazzle readers with were established to mine lead and stories of cave saltpeter playing to obtain animal furs. In 1770. a major role in concluding the War "Long Hunters" explored the middle 188 SALTPETER AND EARLY KENTUCKY HISTORY and southern regions of Kentucky. caves, rock shelters, and the dry They would have been the first to soi 1 under bui 1dings. find the caves later mined for With the Declaration of saltpet-er production. Independence in 1776, France began The Treaty of 1763, ending the supporting the cause of the 13 French and Indian War, gave colonies. firms and supplies for the land between the war were bought from France. Large' fippal achi an Mountains and the quantities of gunpowder were Mississippi River plus Canada. The imported despite the British British discouraged any settlement blockade. The pivotal victory of of their "Northwest Territory," Saratoga in 1777 was made possible which was north of the Ohio River. by gunpowder received from the The Commonwealth of Virginia, French. The Continental Congress which included Kentucky. began appointed a committee to promote surveying its land south of the the manufacture of saltpeter. Ohio River in 1773. As the District committees would purchase movement toward independence from a pound of pure saltpeter for 20 British rule increased, more cents. Numerous caves east of the settlers moved into the Blue Grass Appalachian Mountains were mined Region, despite frequent Indian for saltpeter during and after the attacks. In the spring of 1774, War of Independence. When some of James Harrod built the first cabin the miners moved west, they took in Kentucky, at the town of their knowledge of saltpeter Harrodsburg. production technology with them. In October, 1774, the Rritish The War a+ Independence ended prohibited the export of gunpowder with the defeat of Cornwallis on to the rebellious colonies, and October 17, 1781. The immediate confiscated stores of gunpowder in result of the cessation of Massachusetts. Hostilities hostilities was a surge of increased, and the Revolutionary immigrants into Kentucky. The War began when Eritish Regulars Treaty of Paris in 1783 formally fired on the Minutemen at ended the war, but did not halt Lexington, Massachusetts, on April bitter feelings toward the 19, 1775. British. Enqland- maintained outposts in Indiana and Michigan, At a time when settlers in and supported Indian attacks Kentucky needed saltpeter for against the immigrants. gunpowder to insure their safety, and salt for food, the eastern Saltpeter production remained a communi t ies could not provide it. commercial venture in Virginia The British blockade reauired the after the Revolutionary War. Local colonists to produce their own gunpowder mills produced the essential chemicals. On June 10, explosive for firearms, and to use 1775, the Continental Congress in blasting for land clearing and decreed that all saltpeter the construction. Transport of the colonists possessed had to be explosive over rough roads for delivered to the nearest factory long distances was dangerous. for gunpowder producton. One month Kentucky settlers searched for later, Congress adopted a caves to produce saltpeter for resolution to promote the their own personal needs. Early production of saltpeter within the records do not outline evidence of coloni es. The resolut i on was a commercial saltpeter industry in written by Dr. Benjamin Hush of Kentucky after the end of the Philadelphia, a famed physician Revolutionary War; by 1800, and signer of the Declaration of however, 28 saltpeter caves and Independence. Rush described the rock houses were reported to have methods of production of saltpeter produced 100,000 pounds of from calcium nitrate found in saltpeter. CAVES AND THE SALTPETER INDUSTRY IN KENTUCKY 189

Kentucky achieved statehood on Shepherdsville, in 1779. Three June 1, 1792. Saltpeter production thousand gallons of water were was still a local cottage industry boiled down to yield four bushels requiring on1 y calcium of salt. Immigrant demand for salt nitrate-containing earth, ashes, was so great that at the end of water, digging tools, lumber +or a the War of Independence, the price leaching vat, and a boi 1ing of salt was inflated to over 6500 kettle. By this time, the per bushel. Numerous salt licks saltworks in the State employed were developed in the vicinity of hundreds of men as carpenters, Bullitt's Lick and timber +or wood choppers, boiling kettle fires was cut from a large area. tenders, and waggoners. Salt We1 1s were dug to find salt brine. production began shortly after the The brine was transported to the first settlers entered the State, Soiling kettles by miles of hollow and saltworks were developed wooden pipes made from logs that despite the hazard of Indian were bored by augers, with the attack. Daniel Eoone was captured ends held together by iron bands. by Indians in 1778 while making By the turn of the century, the salt for Boonesboro. The first price of a bushel of salt fell to commercial saltworks was erected 81.00. Figures 1 and 2 show some at Bullitt's Lick, near the of the equipment used in the salt present-day town of and saltpeter industry.

Figure 1. V-leaching vats used in the entrance of Mammoth Cave. From Mammoth Cave Saltpeter Action History Experiment, July 1974. CRF photo, R. Pete Lindsley. 190 SAMUEL BROWN, M.D., AND THE COMMERCIAL SALTPETER INDUSTRY

School in 1799. In 1800, through the influence of Thomas Jef ferson and Dr. Rush, he became a member of the prestigious American Philosophical Society. Brown demonstrated a special interest in industrial chemistry and its application to agriculture. At this time. Lexington had one gunpowder mill which received its saltpeter from nearby caves. In 1801, he visited Kinkaid's Cave (Great Saltpeter Cave) , located 50 miles southeast of Lexington. The cave had been discovered only 2 years earlier by John Baker, but it already supplied much of the saltpeter for Figure 2. One of the original the Lexington mill. Dr. Brown's saltpeter furnace bioling interest in the commercial kettles from Mammoth Cave, production of saltpeter at this recently located near Browns- cave resulted in the most ville, Kentucky. CRF photo, important contemporary article R. Peter Lindsley. published on the subject, "A Description of a Cave on Crooked SAMUEL BROWN, M.D., Creek, with Remarks and Observations on Nitre and AND THE COMMERCIAL Gunpowder." He read his paper SALTPETER INDUSTRY before the American Phi 1osophi cal Dr. Samuel Erown was the first Society at Philadelphia on scientist to describe the process February 7, 1806. of saltpeter production in early Dr. Brown described the process Kentucky. Eorn in Rockbridge of prospecting for the presence of County, Virginia, in 1769, the son "petre dirt" by taste, or the of a Presbyterian minister, he was effacement of foot and handprints raised in the town of Rockbridge. imprinted on the surface of the Nearby was Saltpeter Cave of cave soil. He outlined the process Rockbridge County, close to by which calcium nitrate in the Natural Bridge. Frown attended cave dirt was leached with water medical school in Philadelphia, to yield "mother liquor." Wood and became a private pupil oi Dr. ashes, contining high Benjamin Rush, Chairman of the concentrations of potassium, were Department of Chemistry. as we1 1 leached with water to yield as a physician of great fame. potassium hydroxide or potash. The Undoubted1y , Brown 1earned of Dr. potassi um hydroxide solution was Rush's interest in saltpeter added to the "mother liquor." mining during the Revolutionary yielding potassium nitrate in War. soluton. This solution was Dr. Brown studied under Dr. filtered and concentrated by Rush 3 years. In 1772. he went boiling to make crystalline abroad to study medicine in saltpeter. Dr. Brown, like Dr. Edinburgh, Scotland, receiving his Rush, understood the process by Doctor of Medicine degree in 1797. which the raw material in the cave Upon returning to the United soi 1 became a purif ied chemical , States, he ventured to Lexingon, but could not have understood the Kentucky. and was appointed Head under1ying chemical reactions. The of the Department of Chemistry at element potassium was not isolated Transylvania University Medical as a metal until 1807. CAVES AND THE SALTPETER INDUSTRY IN KENTUCKY 191

Dr. Brown observed that the continued saltpeter mining on a workmen routinely returned the limited scale. In 1904, Dr. Brown leached dirt back to the cave and Thomas Hart purchased the cave f 1oor. Saltpeter production had and increased saltpeter occurred long enough in the cave production, their venture being for the workers to observe that extremely sucressf ul. In 1805, leached soil wouid again develop they widened the scope of their significant concentrations of enterprise and began producing calcium nitrate after 3 to 5 common salt from local springs. In years. Frown thought the process 1906, an anonymous article was could be continued indefinitely, published in the "Medical making the supply of saltpeter Repository," the most popular from a good saltpeter cave medical publication o+ its day. inexhaustible. Entitled "Caverns in Virginia, George Montgomery purchased Kentucky, and Tennessee, which Great Saltpeter Cave in 1802 and Afford an Inexhaustible Supply of Salt-Fetre," it undoubtedly was written by Dr. Frown. With his brief career as a salt and saltpeter producer a success, Brown moved to New Orleans in April, 1806. He did not return to Lexington until 1819 to resume teaching at Transylvania University. Great Saltpeter Cave became the second greatest source of saltpeter in Kentucky during the War of 1812. As many as 60 to 70 men worked at the cave. Each bushel of cave dirt, about 220 pounds, yielded slight1y more than 1 pound of saltpeter (Fig. 3). In 1805, about 150 bushels of cave dirt were processed dai 1y. With improvements on the saltpeter works, it was reported that up to 1,000 pounds of saltpeter were expected to be produced daily. MAMMOTH CAVE AND THE WAR OF 1812 Modern records do not identify the earliest settlers of the Mammoth Cave Region, but James Sturgeon. a Revolutionarv War soldier, filed a military claim on 200 acres east of Mammoth Cave in the fall of 1790. He was probably Figure 3. Saltpetek miners one of the first settlers on the moved large quantities of floor ridges along the south bank of the breakdown to excavate the salt- Green River. Some of these early peter earth. Leaching hoppers settlers were involved in mineral were in the cave pass'age above exploitation as noted by Imlay, the breakdown slabs in the who wrote of Kentucky in 1792: center of the photograph. Dixon "Sulphur is found in several Cave, Mammoth Cave National places in abundance: and nitre is Park. CRF Photo, R. Pete made from earth which is collected Linds ley. from caves and other places to MAMMOTH CAVE AND THE WAR OF 1812 which wet has not penetrated. The them with gunpowder and arms. The making of this salt, in this net result of mounting conflict country, is so common, that many with Great Britain was settlers manufacture their own congressional Declaration of War gunpowder. This earth is on June 18, 1812. discovered in greater plenty on Prior to 1808. Valentine Simons the waters of Green River, than it sold the 200 acres including Dixon is in any other part of Kentucky." Cave and "Big Cave" to John Flatt Today, there is no record of for at least 8116.67. Also prior the earliest saltpeter production to 1808, Flatt assigned the at Mammoth Cave and the Dixon property to brothers John, Cave; however, it is likely that Leonard, and George NcLean for an these were among the first caves unknown sum, but at 1east 8400. In mined in the State. On September January, 1808, 44 acres including 14, 1798, Valentine Simons Dixon Cave were sold to Charles purchased 200 acres for 880, Morton for 8600. Prior to 1910, making a down payment of about the McLeans sold the remaining 156 $10. In 1799, the property was acres including Mammoth Cave to a surveyed and it included "two "saltpeter company" owned by saltpeter caves. " Simons assigned Flemi ng Gatewood and Char 1es rights of the caves to John Flatt, Wilkins for 83,000. and in 1808 Mammoth Cave was known Char1es Wi1 ki ns was a prominent as Flatt 's Cave. Common V-1 eachi ng Lexington, Kentucky salt and vats were used in the entrance of saltpeter merchant. Records Mammoth Cave (Fig. 1) . The extent document that in 1808 he was of the production of saltpeter selling Kentucky saltpeter via from the caves in this period is Phi ladelphi a 5altpeter merchant ~tnknown. Archibald McCall to E.I. du Pont Persisting hostilities with Company of Wilmi ngton , Delaware. Great Britain were soon to have a Wilkins purchased saltpeter made profound effect on nitrate mining at Great Saltpeter Cave, and must and saltpeter production in have known of Dr. Brown's Kentucky. On October 21, 1805, ownership and interest in the Great Britain defeated the French cave. As the British blockade and Spanish in the Battle of tightened prior to the War of Trafalgar and gained control of 1812, sources of saltpeter near the seas. Several weeks later, Lexington could not meet the however, Napolean gained supremacy demand of the large manufacturers on land in Europe. The United of gunpowder such as du Pont. States was the world's Wilkins found his needed sources second-greatest maritime nation at of saltpeter in the caves oi the this time, and was uninvolved Mammoth Cave Region. In 1812, directly in the European war. weal thy Philadelphia merchant and American ships supplied war saltpeter dealer Hyman Gratz materia1 s to Napol eon, prompti nq purchased Fleming Gatewood 's Great Britain to blockade American interest for 810,000. The combined ports. Congress responded with business acumen of Wilkins and passage of the Embargo and Gratz resulted in spectacular Non-Intercourse Act to deprive saltpeter production from Mammoth both France and Great Britain of Cave between 1812 and 1815. American Goods. This action harmed As the price of saltpeter rose United States' trade and sources prior to the War of 1812, the of saltpeter from Spain and India value of caves containing became less accessible. saltpeter a1 so increased. In Kentucky, it was felt that Saltpeter sold for 17 cents a the British were inciting the pound in 1808 with land at Mammoth Indians to attack frontier Cave selling for 814 per acre settlements, and were supplying (Tab1e 1) . When war was declared CAVES AND THE SALTPETER INDUSTRY IN KENTUCKY 193 I Table 1.--Economic Cycle - Mammoth Cave Land and Commodities, 1775- 1084, Land KNOs Gunpowder Corn Year (pound 1 ------(acre) -(pound ----- 1 ------A-A(bushel )

in 1812, saltpeter sold for 70 stores of munitions. Neither Great cents a pound. Hyman Gratz Britain nor the United States purchased his half-interest in gained any clear-cut military Mammoth Cave for 8128 per acre. advantage over the other. This represented an 800% increase The British were exhasted by in the value of the land and cave the Napoleonic struggle; American in 4 years! export trade was para1yzed. The Speculation in Mammoth Cave and British in+luence over the Indians its valuable resource led to a was lost, and the of American fascinating array of maps settlement in the west could not outlining the cave's extensive be stopped. A truce, not saltpeter deposits. Charles surrender, was reached by the Mi1 ki ns ' brother-i n-1 aw, Doctar Treaty of Ghent on December 14. Frederick Ridgley, sent one of the 1814. The large quantities of maps to Philadelphia to Dr. Samuel saltpeter produced in Kentucky did Brown's famous mentor. Dr. not "win the war", but probably Benjamin Rush. Despite lack of a helped to stabilize the survey base and the distortion of war-related inflation in its distances on these maps, they have value. Had the value of saltpeter proven an important source of continued to climb due to information on the early history scarcity, undoubtedly other of the cave. sources of the chemical would have In 1812, Kentucky supplied over been exploited. 300,000 pounds @fsaltpeter from 35 saltpeter caves and rock shelters, with a value of ECONOMIC AFTERMATH approximately 52 mi11 ion. Even OF THE WAR OF 1812 though this represents the bulk of A-fter the war, the value of saltpeter produced in the United saltpeter plummeted. The price of States that year, there is no a pound of saltpeter fell to 15 evidence that this and subsequent cents or less with cessation of production was vital in concluding the blockade and resumption of the War of 1812. The war had no foreign imports. Demand for prolonged battles requiring large saltpeter declined due to its 194 ECONOMIC AFTERMATH OF THE WAR OF 1812 limited usefulness as an Stories of aboriginal remains industrial chemical, and the lack found by saltpeter miners in Short of demand for gunpowder. Transport Cave and Mammoth Cave resulted in of cave saltpeter from the renewed interest in Mammoth Cave. interior of Kentucky to Nahum Ward, of Shrewsbury, Philadelphia was too difficult and Massachusetts, visited the cave in espensive to continue cave October, 1815. On his return to production. As late as 1840, only Massachusetts, he took with him a single land road led to Mammoth one of the "mummies" found in Cave. When the road was wet, Short Cave and displayed in wagons could not reach the cave Mammoth Cave. He wrote a newspaper and visitors traveled by article on his cave trip that was horseback. E. I. du Font continued widely reprinted here and abroad to dominate gunpowder manufacture, for many years. & map of the cave utilizing foreign sources of accompanied his article (Fig. 4). saltpeter . Nahum Ward's account made the cave Economic stagnation and famous, and was crucial to the depression worsened steadi 1y after early commercial tourist activity the war, resulting in the Panic of at Mammoth Cave. 1819. The inflated value of Immediately after saltpeter Mammoth Cave land collapsed (see production ceased at Mammoth Cave Table 1). A1 1 of the saltpeter in the Spring of 1815, the cave caves except Mammoth Cave and became a commercial tourist Great Saltpeter Cave fell into attraction. Thi s commercial iiation disuse. Great Saltpeter Cave was led to preservation of the reported1y mined commerci a1 1y for saltpeter works in the cave, saltpeter during the Mexican War, although surface works were beginning in 1844. destroyed by cave development and

Figure 4. One of the several versions of Nahum Ward's "Map of Mammoth Cave." The Green River is shown over the cave far to the south of its true location. From Sears (1850). CAVES AND THE SALTPETER INDUSTRY IN KENTUCKY 195 natural elements (Fig. 5). The he owned in southern Kentucky. He other saltpeter caves were not so developed an intense interest in fortunate, and their saltpeter Mammoth Cave, that led to the cave works decayed or were destroyed. becoming a world wide tourist Small numbers of workers were used attraction. Dr. Croqhan died in for saltpeter praduction. so no 1847, leaving a will detailing towns or villages were built near future management of the cave. His the caves. Salt praduction estate maintained a control 1ing fluorished after the war, however, interest in the cave until 1727, with producton o+ up to 500,000 when the Mammoth Cave National bushel s annual 1y. Towns bui 1t near Parks Ossociaton purchased a salt licks still exist today. two-thirds interest of the estate. Mammoth Cave and surrounding Mammoth Cave National Park land including 1,614 acres was today protects Mammoth Cave, Dixon purchased for SIO, OOO by Doctor Cave, and three smaller saltpeter John Croghan in 1839. He was an caves, Martin Cave, Jim cave, and enterprising businessman, Long Cave (Grand Avenue Cavern). physician, and farmer, and in Coach Cave and Short Cave, outside addition produced salt from land the park boundary, were mined for

Figure 5. Rotunda saltpeterworks at Mammoth Cave. CRF Photo, R. Pete Linds ley. REFERENCES saltpeter, and are intermittently tion: National Park Service, shown on a commercial basis. Other Cave National Park, 33 p. caves that had limited saltpeter production exist in the Mammoth Faust, Burton, 1955, Saltpeter Cave Region. Only Pruett Saltpeter mining tools used in caves: Na- Cave near Bowling Green had a tional Speleological Society Bu- large-scale operation that would letin, v. 17, p. 8-18. compare with that of Mammoth Cave and Great Saltpeter Cave. Faust, Burton, 1967, Saltpeter Saltpeter Cave in Carter Caves mining in Mammoth Cave, Ken- State Park, and Great Saltpeter tucky: Louisville, Filson Club, Cave also have underground 96 p. saltpeter works preserved, and are open to the public. Hill, Carol A., ed., 1981, Salt- peter: National Spel eol ogi cal In 1973, the Saltpeter Group of Society Bulletin, v. 43, 48 p. the Cave Research Foundation addressed the challenges of the Hill, Carol A., 1982, Saltpeter science and history of cave caves of the United States - nitrate formation and saltpeter updated 1i st: National Spel eol o- production. Under the leadership gical Society Bulletin, v. 11, of Carol A. Hill and Duane p. 24-27. DePaepe, the results of their studies were published in the Jackson, George F., 1949, Salt- "Bulletin of the National peter mining in American caves: Speleological Society in October, National Speleological Society 1981. This remains the definitive Bulletin, v. 11, p. 24-27. study of the scientific aspects of cave saltpeter. McDowel 1, Robert E. , 1956, Bul- litt's Lick, the related salt- hcknowl edgements: Harold Meloy , works and settlements: Filson Alfred Scheide, and Richard A. Club History Quarterly, v. 30, Watson provided their insight and p. 241-269,. suggestions for this chapter. Meloy, Harold, 1984, Mummies of REFERENCES Mammoth Cave: She1 byvi 11e, Indi- ana, Micron Publishing Company, The purpose of the following 43 p. references is to provide recent accessible references f rom White, Wayne R., 1967, The Speleo- speleological 1i terature. These graphy of Great Saltpeter Cave: provide access to the more National Spel eol ogical Society extensive bibliography used in News, v. 25, p. 169. writing this chapter. DePaepe, Duane, 1979, The salt- Wigginton, Eliot, ed., 1979, Fox- peter era at Mammoth Cave, a fire 5: Garden City, New York, cultural resources investiga- Anchor Books, p. 246-260.