4 Reaction Mechanisms Part (i)Pericyclic Reactions By R. J. BUSHBY Department of Organic Chemistry The University Leeds LS2 9JT A new two-volume work on pericyclic reactions has been published’ which contains a number of useful articles particularly those where a fresh view of the literature has been obtained by dividing according to substrate type (carbanions carbonium ions etc.) rather than the more normal division by reaction type. 1 Cycloadditions and Cycloreversions More energy surface calculations on cycloaddition reactions have been reported2 and in an interesting paper Houk3 has compared the results of semi-empirical and of ab initio MO methods. For the Diels-Alder reaction and the addition of 1,3-dipoles to olefins he concludes that most NDO semi-empirical methods favour an unsymmetrical (biradical) mode of addition whereas ab initio methods favour a symmetrical (concerted) transition state (TS) because neglect of overlap even in the MIND0 method results in the filled-orbital-filled-orbital repulsions being under-estimated.In particular these NDO methods under-estimate the repulsive interaction between the two highest occupied molecular orbitals which for an allowed cycloaddition contributes more to the destabilization of the unsymmetrical TS than to the symmetrical TS. Although this provides a simple explanation of why the results obtained by the two methods differ it does not enable one to judge which is closer to the truth as the effect of correlation on the balance between symmetric and unsymmetric TS remains largely a matter for conjecture.However the sym- metric TS seems the more likely as Salem’s work on the Diels-Alder reaction had previously shown that a parallelism exists between the results of STO-3G and 4-31G +3 x 3CI calc~lations.~ As far as the experimental evidence on [2+4] cycloaddition reactions is con- cerned Firestone has produced another lively article5 which contends that this points towards biradical intermediates but most authors seem to favour the more balanced arguments forwarded by Huisgen in favour of the concerted mechanism.6 ‘Pericyclic Reactions’ ed. A. P. Marchand and R. E. Lehr Academic Press New York 1977 Vols. I and 11. (a) L. A. Burke and G. Leroy Theor. Chim. Acta 1977 44 219; cf. ibid. 1975 40 313; (b)M.V. Basilevsky A. G. Shamov and V.A. Tikhomirov I. Amer. Chem. SOC.,1977,99 1369. P. Caramella K. N. Houk,and L. N. Domelsmith J. Amer. Chem. Suc. 1977,99,4511. R. E. Townsend G. Ramunni G. Segal W. J. Hehre and L. Salem J. Amer. Chem. Soc. 1976 98 2190. ’ R. A. Firestone Tetrahedron 1977,33 3009. ‘R. Huisgen J. Org. Chem. 1976,41,403. 59 R. J. Bushby There can be no doubt however that suitable substitution can divert most of these reactions into a two-step pathway and also that intermediates are involved in many thermal 12+21 cycloaddition reactions. In an interesting discussion of the nature of these intermediates and of the problems involved in their interception7 it has once again been stressed that an over-rigid distinction cannot be drawn between the biradical (1)and the zwitterion (2).However the reaction of tetracyanoethylene with enol ethers has been taken as a model of one involving an intermediate of high zwitterionic type.76 a-b-a-b + (1) (2) Most treatments of orientation in cycloaddition reactions have been in terms of PMO theory' and Trong Anh and co-workers have pointed out the advantages of their simple approach to the problem.8u*b This concentrates on selecting the site at which bonding is first likely to occur by looking for the most important HOMO- LUMO two -cenPre interaction (rather than the four-centre interaction normally employed) which then dictates the course of the reaction. This method produces reliable predictions in most cases even the troublesome case of acrolein dimeriza- tion," and the results are furthermore independent of the degree of concertedness of the reaction.As well as the more traditional cr-p correlations,' localization" and FMO" approaches have also proved their value in rationalizing relative reactivities. Hence the energy barriers for the cycloaddition of maleic anhydride to various aromatic systems may be correlated with the corresponding changes in resonance energy." On the other hand as predicted by FMO theory values of logk for the Diels-Alder reaction of various l-aryl-4-benzylidene-5-pyrazolones""(3) Ph Ar (3) correlate well with their measured reduction potentials and the rates of addition of diazomethane to double bonds correlate with experimental estimates of the energy difference E(HOM0 diazomethane)- E(LUM0 dipolarophile.l16 When '(a)R.Huisgen Accounts Chem. Res. 1977 10 199; (6) ibid. p. 117. (a)0.Eisenstein J. M. Lefour N. Trong Anh and R. F. Hudson Tetrahedron 1977 33 523; (b)C. Minot and N. Trong Anh ibid. p. 533; (c) V. Bachler and F. Mark ibid. p. 2857; (d)V. Bachler and F. Mark Theor. Chim. Acta 1976 43 121. C. K. Bradsher T. G. Wallis I. J. Westerman and N. A. Porter J. Amer. Chem. SOC.,1977,99,2588;S. Ito and I. Saito Tetrahedron Letters 1977 1203. lo W. C. Herndon J.C.S. Chem. Comm. 1977 817; cfi R. D. Brown J. Chem. SOC. 1950,691 2730. '' (a)G. Desimoni P. P. Righetti E. Selva G. Tacconi V. Riganti and M. Specchiarello Tetrahedron 1977,33 2829; (b)J. Geittner R. Huisgen and R. Sustmann Tetrahedron Letters 1977,881;(c)ibid.p. 877. Reaction Mechanisms-Part (ii) Pericyclic Reactions 61 however CNDO calculated energies are employed more than just the simple FMO interactions have to be included.”“ Further kinetic evidence has been provided for a twisted quasi-zwitterionic TS in the thermal retro-[2 +21 reactions12 of cyclobutanones and among photochemical [2+21 reactions a particularly neat example has been reported by Paq~ette’~ in the conversion of compounds (4) into (5) via the caged intermediate (6). Me Me Me Studies of the Diels-Alder reaction have once again emphasized that factors besides secondary orbital interactions particularly steric factors can influence endolexo ratios14 and X-ray crystallography has confirmed that the adduct formed between 2-cyano- 1,3-diphenylallyl anion and trans-stdbene also has the expected geometry (7) for a concerted [2 +4] suprafacial addition process.1s Among the related [2 +41 addition reactions of 1,3-dipoles Huisgen has reported details of work on azomethine imines16 and Padwa has continued his studies of nitrillium ylides.” He has suggested”” that the observation that ylide (8; R = H CF3 or aryl) gives adduct (9) whereas the corresponding ylide in which R is alkyl gives adduct (lo) can be rationalized by consideration of the geometry of the intermediate.” Hence the intermediate with R=H or aryl is thought to be linear R RYR ph+rph 12 H.M. Frey and R. A. Smith J.C.S. Perkin II 1977 752; ibid. p. 2082. 13 L. A.Paquette T. G. Wallis K. Hirotsu and J. Clardy J. Amer. Chem. Soc. 1977,99 2815. 14 T. Sasaki K. Kanematsu K. Iizuka and N. Izumichi Tetrahedron 1976,32,2879;D. W. Jones J.C.S. Perkin I 1977 980; K. Seguchi A. Sera and K. Maruyama Bull. Chem. SOC.Japan 1976 49 3558. 15 W. T. Ford and G. F. Luteri J. Amer. Chem. SOC.,1977,99 5330. 16 R. Huisgen and A. Eckell Chem. Ber. 1977,110,522,540,559,571;R. Huisgen M. V. George A. S. Kende and A. Eckell ibid. p. 578. 17 (a) A. Padwa P. H. J. Carlsen and A. Ku J. Amer. Chem. SOC.,1977 99 2798; (b) A. Padwa and P. H. J. Carlsen ibid. p. 1514; A. Padwa and N. Kamigata ibid.,p. 1871. P. Caramella R. W. Gandour J. A. Hall C. S. Deville and K. N. Houk J. Amer. Gem. Soc. 1977,99 18 385. R.J. Bushby and to behave as a 1,3-dipole but the intermediate with R=alkyl is thought to be bent with high resultant carbene character. An independent synthesis and study of the reactions of compound (11)l9 has shown that of the two adducts formed between tropone and cyclopentadiene (12) and (13) the first probably arises by initial formation of (11) followed by a Cope (11) (12) (13) rearrangement but that the second cannot be formed from this intermediate. It is probably the result of a concerted [4 +61 cycloaddition. This completes an interes- ting analogy between this reaction and the reaction of cycloheptatriene and 2,5-dimethyl-3,4-diphenylcyclopentadienone.20 Whereas it has now been proved that the reported [2+6] adduct between N-ethoxyazepine and diethyl azodicarboxylate is in fact a normal Diels-Alder product2'" the formation of an analogous adduct (14) with nitrosobenzene has been confirmed by X-ray a,:'ph NCO2Et (14) The reaction of electron-deficient acetylenic dienophiles with the pentalenes (1 5) (X =H CHO CN or C02Me) has been shown to involve attack on the electron- rich ring to give the [2+8] adducts (16) whereas cyclopentadiene attacks the electron-poor ring to give the [2+4] adducts (17).22 It is also interesting to note (15) (16) (17) that although 8-phenylheptafulvene seems to give with tetracyanoethylene the expected [2+8] adduct (18) the 8,8-diphenyl compound gives a [2+4] adduct l9 M.Franck-Neumann and D. Martina Tetrahedron Letters 1977 2293. 2o K.N. Houk and R. B. Woodward J. Amer. Chem. SOC.,1970,92,4143. '' (a)W. S. Murphy and K. P. Raman J.C.S. Perkin I 1977,1824; (6)W. S. Murphy K. P. Raman and B. Hathaway ibid. 2521. 22 M. Suda and K. Hafner Tetrahedron Letters 1977,2449,2453; see also B. Kitschke and H. J. Lindner ibid. p. 25 11. Reaction Mechanisms-Part (ii) Pericyclic Reactions (19).23 In this case presumably steric rather than electronic factors are responsible for the change in selectivity. A remarkable formal [2+8] adduct (20) whose structure has been confirmed by X-ray crystallography is also formed in the reaction of 4-phenyl-1,2,4-triazoline-3,5-dione with ~ctalene.~~ Ph I (18) (19) (20) Aspects of the elimination of nitrogent5 and of sulphur dioxide26 have been reviewed and a detailed study of the elimination of nitrogen from optically active cis- and trans-3-ethyl-5-methylpyra~olines~~ has shown that the results for the two isomers differ so that no universal stereochemical course can be assigned to these reactions.It has also been shown that elimination of nitrogen is enormously accelerated by the introduction of a trimethylsilyl group in the 4-po~ition.~~ This has been rationalized in terms of concerted nitrogen loss and silyl migration (21). Hammett LFER studies of the elimination of sulphur dioxide from the thiiren dioxides (22) have provided evidence for a stepwise process29 and it has been shown that elimination of sulphur monoxide from the cis- and trans-thiiran oxide (23) proceeds with 95% retention of stereo~hemistry,~’ in contrast to earlier results for related systems.306 2 SigmatropicReactions STO-3G and 4-31G ab initio MO calculations on the 1,3-sigmatropic rearrange- ment of vinyl alcohol suggest that the hydrogen follows an antarafacial route but that the energy barrier to this is high (calculated as 85 kcal m~l-’).~’ This apparent 23 K.Komatsu M. Fujimori and K. Okamoto Tetrahedron 1977 33 2791. 24 E. Vogel H.-V. Runzheimer F. Hogrefe B. Baasner and J. Lex Angew. Chem. Internat. Edn. 1977 16 871. ’’ H. Meier and K.-P. Zeller Angew. Chem. Internat. Edn. 1977 16 835. 26 W. L. Mock,in Vol. 11of ref. 1. 27 T. C. Clarke L. A. Wendling and R. G. Bergman J. Amer. Chem. SOC.,1977,99,2740. 28 R. F. Cunico and H. M. Lee J. Amer. Chem.Soc. 1977,99,7613. 29 J. C. Philips and 0.Morales J.C.S. Chem. Comm. 1977 713. 30 (a) W. G. L. Aalbersberg and K. P. C. Vollhardt J. Amer. Chem. Soc. 1977 99 2792; (6) G. E. Hartzell and J. N. Paige J. Org. Chem. 1967 32 459; K. Kondo M. Matsumoto and A. Negishi Tetrahedron Letters 1972 2131. 31 W. J. Bouma D. Poppinger and L. Radom J. Amer. Chem. SOC.,1977,99,6443. R. J. Bushby resistance to unimolecular tautomerism has been advanced as an explanation of the stability of this species in the vapour phase. In the liquid form tautomerism is rapid but here bimolecular mechanisms are far more likely. In contrast semi-empirical MO calculations on other 1,3-sigmatropic shifts predict a suprafacial route prob- ably with high biradical/zwitterionic character,32 and experimental evidence in support of this latter prediction has been 1,5-Sigmatropic shifts in indenyl systems continue to attract attention,34 and for the rearrangement of the vinyl-substituted indenes (24) it has been shown that a good LFER exists between the rate of the sigmatropic shift and that for the corresponding attack of morpholine on CH2=CHX.34" This emphasizes the importance of the secondary orbital interaction between the HOMO(indeny1) and LUMO(migrating vinyl group).A similar secondary orbital interaction (25) has been used to explain the high rate of rearrangement of the spirononatriene (26) relative to its isomer (27) (rate ratio > 1000 l).35Studies of conformationally constrained indenyl suggest that the migrating vinyl group prefers an em orientation as in compound (28) rather than an endo orientation as in compound (29).An exu orientation avoids the unfavourable secondary orbital interactions (marked *) found in the endo TS (30). LUMO ethylene CH=CHX HOMO pentadienyl W. W. Schoeller J. Amer. Chem. SOC.,1977 99 5919; J. P. Grima F. Chopfin and G.Xaufmann J. Organometallic Chem. 1977,124 315; T. Minato S. Inagaki H. Fujimoto and K. Fukui Bull. Chem. SOC.Japan 1977,50 1651. and J. E. Baldwin S. E. Branz and J. A. Walker J. Org. Chem. 1977 42 4142d W. Kirmse and H.-R. Murawski J.C.S. Chem. Comm. 1977 122. (a)D. J. Field and D. W. Jones J.C.S. Chem. Comm. 1977,688;(b)D.W. Jones and G. Kneen J.C.S. Perkin I 1977 1313; W. A. Pettit and J. W. Wilson J.Amer. Chem. SOC.,1977 99 6372; K. K. de Fonseka C. Manning J. J. McCullough and A. J. Yarwood ibid.,p. 8257. M. F. Semmelhack H. N. Weller and J. S. Foos J. Amer. Chem. SOC.,1977,99,292. Reaction Mechanisms-Part (ii) Pericyclic Reactions Schmid and co-~orkers~~ have provided impressive evidence for potential 2,5 -biradical character (31) in the TS of the Cope rearrangement by studying the quantitative effect of radical-stabilizing substituents at these positions uiz. (32; X I X (31) (32) R=Me X=H CN or C02Me). They point out however that this biradical mechanism is unable to explain all of the phenomena associated with these re- arrangements and suggest that they are normally concerted and only diverted into the biradical pathway when such 2,5-substituents are present.Similar studies have been reported by Dewar3’ who however no favours this flexible mechanism hypothesis but proposes a 2,5-biradicaloid even in the rearrangement of hexa-1,5-diene. His MIND0/3 calculation^^^ suggest that this is a true inter- mediate and that it prefers a chair conformation. The balance between boat and chair TS (or intermediate!) in other 3,3-sigmatropic rearrangements has been investigated by several groups.39 For 3,3’-dimethyl-3,3‘-bi~yclopropenyl~~~ the chair-like TS (33) is preferred over the boat-like TS (34) by 4.3 kcal mol-’ and a chair-like TS also seems to be preferred in the thermal (35a) and acid-catalysed (35b) amino-Claisen rearrangement^.^^^ These latter reactions have been studied in some and appear to involve a concerted pericyclic mechanism.The large acceleration brought about by protonation of the nitrogen is attributed to charge delocalization in the reaction (36). In contrast the Claisen rearrangement (35) a;X=NH b;X=NHz+ (3 ) implicated in t,,e reaction of allylic alcohols with cyclic orthoe~ters~~~ may proceed either predominantly through a chair-like TS or a boat-like TS depending largely on steric factors and the two TS’s also seem to be finely balanced in the rearrangements of ortho-dienones such as compound (38).39d In this particular 36 (a)R. Wehrli H. Schmid D. Bellus and H.-J. Hansen Helu. Chim. Actu 1977 60 1325; full English summary Chimia (Swirl.),1976,30,416. 37 (a)M. J. S. Dewar and L. E. Wade J.Amer. Chem. Soc. 1977,99,4417;(6) cf. J. Amer. Chem. Soc. 1973 95 290. M. J. S. Dewar G. P. Ford M. L. McKee H. S. Rzepa and L. E. Wade J. Amer. Chem. Soc. 1977,99 5069. 39 (a)J. H. Davis K. J. Shea and R. G. Bergman J. Amer. Chem. Soc. 1977,99,1499; (b)S. Jolidon and H.-J. Hansen Helv. Chim. Actu 1977 60 978; (c) R. J. Cave B. Lythgoe D. A. Metcalfe and I. Waterhouse J.C.S. Perkin I 1977 1218; (d)A. Wunderli T. Winkler and H.-J. Hansen Helv. Chim. Actu 1977,60 2436. R. J. Bushby case the threo-form rearranges mainly via a chair-like TS (39) whereas the erythro-form prefers a boat-like TS (40). (37) (38) (39) (40) Several studies have been reported which bear on the possible involvement of radical pairs in the sigmatropic rearrangements of anionic and zwitterionic Hence the 2,3-sigmatropic rearrangement of the anion (41) normally proceeds with inversion of the ally1 group but as the temperature of the reaction is raised there is some scrambling of the isotopic label (marked *) pointing to a dissociation-recombination mechanism.40a Studies of pentadienyl-N-acyl- ammonioamidates (42)"Ob show that the proportions of the 1,2- and 5,2-rear- rangement products (43) and (44) are characteristic only of the substituents R1,R2 and R3and not the group X and hence these rearrangements are formulated as involving the radical intermediate (45) whereas the proportion of the 3,2-re- arrangement product (46) is a function of the nature of X and this product must be formed at least in part by a different mechanism possibly via the concerted TS (47).Studies of the Wittig rearrangement of the optically active ether (48)40c show similar degrees of racemization in the products of both 1,2- and 1,4-shifts in contrast to the corresponding Stevens rearrangement where the product of the Me i-NCOX Me,N-NCOX R:;+yyy 3 Me2N-NCOX 4 R2 R3 R2 R3 R2 R3 (42) (43) (44) Me,N-"COX Me,?-yNCOX R' R' / R2 R3 R2 R3 R2 R3 40 (a)E. Grovenstein and A. B. Cottingham J. Amer. Chem. SOC. 1977 99 1881; (b) K. Chan-trapromma W. D. Ollis and I. 0. Sutherland J.C.S. Chem. Comm. 1977 97; (c) H. Felkin and C. Frajerman Tetrahedron Letters 1977,3485; (d)E. F. Jenny and J. Druey Angew. Chem. Internat. Edn. 1962 1 155. Reaction Mechanisms-Part (ii) Pericyclic Reactions 1,4-shift is the more highly racemized of the (Scheme).An attractive simple explanation of this observation is based on the assumption that the anion (48) adopts a cis configuration and the anion (50) an all-trans configuration; then the radical pair (49) is presumably generated in the configuration shown in which the a-phenylethyl radical has an equal distance to travel to form either product whereas the radical pair (51) is generated in an extended configuration where it hzs much further to travel in order to produce the 1,4-rearrangement product with consequently greater chances for racemization. Ph P -H -++Me + p0-CLPh 0-Ph Me / Me -30% racemization -30% racemization Ph Me Me + PhL N M % NMez -0% racemization -50% racemization Scheme Various mechanisms have been demonstrated for dyotropic rearrangements of silicon compounds all of which appear to involve some degree of nucleophilic attack on silicon.41 Hence the rearrangement of the allyl system (52) is thought to involve initial co-ordination between the oxygen and the silicon followed by migration of the allyl group as shown (53),41" and the fluoride-ion-catalysed rear- rangement of the same compound to proceed via initial displacement of the trimethylsilyl group to give the intermediate (54).41b Similarly the rearrangement of the ester (55) is formulated as involving the intermediate (56).41c 41 (a)M.T. Reetz Chem. Ber. 1977 110 965; (b) M. T. Reetz and N. Greif ibid. p. 2958; (c) M. T. Reetz and N.Greif Angew. Chem. Infernat.Edn. 1977 16 712. R. J. Bushby Me Me I +- Ph2C- SiMe,I o+c,o I I R R (55) (56) 3 Ene Reactions Detailed studies of the ene reaction of terminal alkenes with maleic anhydride have been interpreted in terms of a concerted mechanism and a preferred ex0 TS,42 and the stereochemistry of the dimer formed from the tricycloundecene (57) has also been advanced as evidence for a concerted Lewis-acid catalysis of the ene reaction of chloral allows the reaction to be performed under mild synthetically useful conditions and added ferric chloride has also been shown to reverse the preferred stereochemistry in the reaction with a-~inene.~~ This can be understood in terms of the simple steric model illustrated in formulae (58) and (59) I H (57) (58) .(59) in which it is assumed that the reactions are concerted and that in the ferric chloride-chloral complex the trichloromethyl and ferric chloride groupings are trans to each other. 4 Electrocyclic Reactions Several theoretical studies of electrocyclic reactions have been reported45 and calculations of the energy surface for the opening of cyclopropylidenes to allenes have revealed a complex combination of disrotatory and conrotatory motions and provided an explanation of how chirality is retained.45a 42 F. R. Benn J. Dwyer and I. Chappell J.C.S. Perkin 11,1977 533. 43 R. Greenhouse W. T. Borden T. Ravindranathan K. Hirotsu and J. Clardy J. Amer. Chem. Soc. 1977,99,6955. 44 G. B.Gill and B. Wallace J.C.S. Chem. Comm. 1977 380 382. 45 (a) P. W. Dillon and G. R. Underwood J. Amer. Chem. Soc. 1977 99 2435; (b) N. L. Bauld and J. Cessac ibid. p. 23; Y.Jean and A. Devaquet ibid. p. 1949; M. J. S. Dewar G. P. Ford and H. S. Rzepa J.C.S. Chem. Comm. 1977,728; R. Cimiraglia,M. Persico and J. Tomasi J. Phys. Chem. 1977 81,1876. Reaction Mechanisms-Part (ii) Pericyclic Reactions The ring-opening reactions of Dewar-ben~ene"~ have and of bi~yclopentene~~ been shown to proceed in a straightforward manner even though formally disallowed. In the case of 5 -methylbicyclopentene Bald~in"~ has provided evi- dence that products (60) and (61) arise by a 'hot molecule' rearrangement of the initial 1,4-bond cleavage product as a result of which he claims that the contro- versy which has surrounded these rearrangements may be finally 'retired'.The electrocyclic ring-opening reactions of ep~xides"~ have been reviewed48a and trapping of the resultant carbonyl ylides has shown that as expected the thermal reaction proceeds mainly in a ~onrotatory~~~*~ and the photochemical reaction mainly in a disrotatory manner.48d7e In both cases however some cross- over products are obtained and these seem to arise not only by isomerization of the intermediate but also by a competing ring opening in the disallowed sense. In the case of vinyl-substituted sy~terns,~~~'~ such as compound (62),48' the initial product (63) of conrotatory ring opening can be trapped with maleic anhydride or may undergo a disrotatory closure to form a five-membered ring (64).It is interesting to note that n.m.r. observation of the reduction of the cyclopropyl bromide (65) with lithium in THF at -60 "Cshows initial formation of the 2,E-anion (66)49 which then more slowly isomerizes to the equilibrium Z,E/E,E Ph Ph mixture. This is consistent with the expected disrotatory opening of an inter- mediate cyclopropyl radical.49b Attempts to probe the stereochemistry of the 46 M. J. Goldstein and R. S. Leight J. Amer. Chem. Soc. 1977,99 8112. " G. D. Andrews and J. E. Baldwin J. Amer. Chem. SOC.,1977,99,4853. (a)R. Huisgen Angew. Chem. Internal. Edn. 1977 16 572; (6) V. Markowski and R. Huisgen J.C.S. Chem. Comm. 1977,439,440; (c) M. S. Medimagh and J. Chuche Tetrahedron Letters 1977,793;(d) W.Eberbach and U. Trostman ibid. p. 3569; (e)V. Markowski and R. Huisgen ibid. 1976 p. 4643. 49 (a) G. Boche and D. R. Schneider Tetrahedron Lerters 1976 3657; (b) cf. M. Newcomb and W. T. Ford J. Amer. Chem. SOC.,1974,96 2968; and S. Sustmann and C. Ruchardt Chem. Ber. 1975.108 3043. R. J. Bushby related pentadienyl anion-cyclopentenyl anion interconver~ion~~ have been largely frustrated by uncertainties which surround the geometry of the reacting or by isomerization of the reaction produ~t.”~ Thermal cyclization of the 12welectron system (67) gives the expected product of conrotatory closure (68).’l This cannot however be claimed as a great success for the predictive power of the orbital symmetry rules as the reaction is reversed photochemically ! (a)D.H. Hunter S. K. Sim and P. R. Steiner Canad. J. Chem. 1977 55 1229; (b)C. W. Schoppee and G. N. Henderson J.C.S. Perkin I 1977 1028. H. Sauter B. Gallenkamp. and H. Prinzbach Chern.Ber. 1977,110 1382.